Skip to content · Gamingdeputy Japan · Nvidia GeForce 572.60 WHQLドライバーはダウンロードできます · February 27, 2025 by Lee Michaelis · Nvidiaは、グラフィックカード用の新しいドライバーをリリースしました。グリーンチームによると、Nvidia ...

Nvidia GeForce 572.60 WHQLドライバーはダウンロードできます - Gamingdeputy Japan

Skip to content · Appomattox County Public Schools Reopen After Investigating Cybersecurity Incident · Published: Feb. 23, 2025 at 6:47 PM EST|Updated: 3 hours ago · Schools · Appomattox County Cybersecurity Incident · Updated: Feb. 21, 2025 at 11:18 PM EST · Here @ Home · YMCA Combat ...

Appomattox County Public Schools Reopen After Investigating Cybersecurity Incident

Skip cloud storage fees forever with 2TB of lifetime space, AES 256-bit encryption, easy file sharing, and no speed limits—just a one-time $89 payment.

Ditch storage limits with a 2TB lifetime cloud storage plan - Tech Advisor

Skip to content · Man Arrested in Gaming Store Robbery · Published: Feb. 8, 2025 at 6:12 PM EST|Updated: 3 hours ago · WDBJ First Alert Hometown Weather: Saturday evening update · Updated: 1 hour ago · Business News · Long Way Brewery in Radford · Updated: 3 hours ago · Safety · Lynchburg ...

Man Arrested in Gaming Store Robbery

It's the latter front, the Multi Frame Generation, that offers the real allure of the new GPU, though. That's because in the games where it's enabled, in many cases at 4K or 1440p, you're talking 4x the frame rate of native res. Case and point, going from 20 fps to 130 fps in Cyberpunk 2077.

Got the Nvidia 50-series and AMD X3D stock-out blues? Skip the waiting lists with this surprisingly well-priced RTX 5080 and Ryzen 7 9800X3D gaming PC | PC Gamer

Skip to main content · Microsoft · Source Asia · Cancel · India · Category: AI · Microsoft AI CEO Mustafa Suleyman says AI companions will create a new class of digital experience

Microsoft announces US $3bn investment over two years in India cloud and AI infrastructure to accelerate adoption of AI, skilling and innovation

Skip to content · Gamingdeputy Japan · NVIDIA、バーチャファイター 6 の最初のゲームプレイビデオを公開 · January 7, 2025 by Lee Michaelis · CES 2025 で、NVIDIA はバーチャ ファイター 6 の最初のゲームプレイ ビデオを公開しました。そ...

NVIDIA、バーチャファイター 6 の最初のゲームプレイビデオを公開 - Gamingdeputy Japan

Skip to content · Gamingdeputy Japan · NVIDIA アプリは PC ゲームのパフォーマンスを低下させますか · December 16, 2024 by Lee Michaelis · 物語 浮上した 今週末、NVIDIA アプリが PC ゲームのパフォーマンスを低下させる可能性があること...

NVIDIA アプリは PC ゲームのパフォーマンスを低下させますか? - Gamingdeputy Japan

It's the final week before the fantasy football playoffs in many leagues. Matt Harmon breaks down what to watch for.

Binge, Stream, Skip: Week 14 Fantasy Football - Yahoo Sports

Skip

skip

Authors:
(1) David Raposo, Google DeepMind and with equal contribution;
(2) Sam Ritter, Google DeepMind;
(3) Blake Richards, Google DeepMind and McGill University & Mila;
(4) Timothy Lillicrap, Google DeepMind;
(5) Peter Conway Humphreys, Google DeepMind;
(6) Adam Santoro, Google DeepMind and with equal contribution. Editor's note: this is part 4 of 5 of a study detailing a way to make transformer-based language models more efficient by dynamically allocating computational resources. Read the rest below. Table of Links Introduction
Background
Implementing Mixture-of-Depths Transformers


3.1. Defining a compute budget


3.2. Routing around transformer blocks


3.3. Routing schemes


3.4. Routing implementation


3.5. Sampling and 3.6. Training methods



Results

4.1. Training, isoFLOP comparisons
4.2. Auto-regressive Evaluation and 4.3. Mixture-of-Depths-and-Experts (MoDE)


Discussion and References 4. Results 4.1. Training, isoFLOP comparisons We first trained models with a relatively small FLOP budget (6e18) to determine optimal hyperparameters (see figure 3). In general, we found that MoD transformers drag the baseline isoFLOP curve "down and to the right". That is, the optimal MoD transformer achieves a lower loss than the optimal baseline, and also has more parameters. A fortunate consequence of this effect is that there exist smaller MoD models that, while they are not themselves isoFLOP optimal for their hyperparameter setting, are nevertheless as- or better-performing than the optimal baseline model while being faster to step. For example, a 220M parameter MoD (figure 3 model #3) variant slightly outperforms the isoFLOP optimal baseline (also 220M, figure 3 model #1), but is upwards of 60% faster to step during training. Crucially, when run on equivalent hardware these two model variants take take approximately the same amount of wall-clock time to train (figure 3). We tested routing every block or every other block, using capacities from 12.5% to 95% of the total sequence. While routing every other block was crucial for strong performance, we found that aggressive capacity reduction was best (gradual improvements were observed when reducing the capacity down to 12.5% of the total sequence, corresponding to 87.5% of tokens routing around blocks, with performance degrading beyond this point). So, it seems the network is robust to significant capacity reductions as long as there is frequent opportunity for full capacity self-attention and MLP computations. Learned routing is crucial, as MoD transformers that use stochastic routing (implemented using a top-𝑘 operation on router weights sampled from a Gaussian distribution) perform drastically worse than both the baseline and normal MoD transformer (figure 3). Depicted in figure 4 is an isoFLOP analysis for 6e18, 2e19, and 1e20 total FLOPs. The trend that FLOP-optimal MoD transformers have more parameters than the baseline continues for these larger FLOP budgets. Notably, there exist MoD variants that are appreciably faster to step than the isoFLOP-optimal baseline (measured as steps per second when training on equivalent hardware) while also achieving a lower loss (in figure 4 we depict normalized FLOPs per forward pass rather than wall-clock step time per se, but from our experiments the two are tightly correlated. A similar plot can be produced showing relative wall-clock step times and the same basic trend is present). Step-wise speed gains come from two sources. First, the FLOP-per-parameter ratio in MoD transformers is less than in the baselines because some proportion of tokens are routed around blocks. So, for a given model size, a transformer requires fewer FLOPs per forward pass. Second, since isoFLOP-optimal MoD transformers are both bigger and achieve a lower loss than the isoFLOP-optimal baseline, there exist smaller MoD variants that perform as well or better than the isoFLOP-optimal baseline, and these variants are faster to step because they are smaller. Altogether, then, there exist MoD transformers that perform as well as isoFLOP-optimal baselines and are faster to step, both because they use fewer FLOPs per parameter and because they use fewer parameters. Figure 4 also reveals another important finding: the optimal MoD transformer is that which uses as many FLOPs per forward pass as the isoFLOP optimal baseline. This finding allows one to directly predict which sized MoD transformer will perform optimally for a given isoFLOP training budget: one just needs to tune the model size for a given MoD configuration (i.e., capacity and routing frequency) to produce a model that uses as many FLOPs per forward pass as the isoFLOP-optimal baseline, and they will have the optimally performing MoD variant for that configuration. Empirically, we find that it is better to add depth than to add width when adding FLOPs to the model. Nevertheless, while the FLOPs per forward pass determines which model will be the isoFLOP optimal, it does not predict whether the optimal loss will improve upon the baseline (see figure 3. Namely, the optimal capacity appears to be empirically determinable. We found that it is best to use 12.5% capacity blocks, every other block. We noticed that MoD transformers had memory savings relative to equivalently sized baseline models at larger sizes, with some variants requiring fewer total devices (i.e., a smaller TPU topology). We did not study this extensively, but we anticipate that as one scales to larger models, these savings could be an important consideration when choosing model variants to train, and could have significant positive effects in regards to the KV cache size during autoregressive sampling. Figure 5 shows the routing decisions for an MoD transformer trained with interleaved routing blocks. Despite aggressive routing around the blocks, transformers are able to achieve performance improvements relative to baselines. We observe patterns that might warrant further study; namely, some tokens appear to engage each block along the transformer’s depth, while others decide to route around blocks whenever possible. Preliminary analyses suggest that the tokens that engage with blocks more frequently are correlated with output predictions that have higher entropy, which possibly corresponds to predictions that are more difficult to make. 4.2. Auto-regressive Evaluation We evaluated MoD variants during auto-regressive sampling (see figure 6). Each model was tested on exactly the same held-out data comprising 256000 sequences (500M tokens). When switching from the top-𝑘 routing method to the predictor-based routing method we observed little performance degradation. As in the training setting, there exist MoD variants that are better performing than the isoFLOP-optimal baseline, while requiring fewer FLOPs per forward pass. These results suggest that the compute savings offered by MoD transformers should translate beyond the training setting. 4.3. Mixture-of-Depths-and-Experts (MoDE) The MoD technique can be naturally integrated with MoE models (together comprising MoDE models) in addition to vanilla transformers. In figure 7 we present results showing that the performance improvments offered by MoD compound with those of MoE. We tried two variants: in staged MoDE, which routes tokens around or towards blocks prior to the self-attention step, and integrated MoDE, which implements MoD routing by integrating “no-op” experts among the conventional MLP experts. The former is advantageous because it allows for tokens to skip the self-attention step, while the latter is advantageous because it simplifies the routing machinery. We noticed that implementing MoDE in the integrated manner was distinctly better than simply reducing the capacity of experts in conventional MoE models, and relying on token dropping to implement residual routing. We believe this is because with the integrated MoDE machinery, tokens explicitly learn to choose the residual path around the experts, as opposed to preferring an expert but being dropped when implemented as a capacity reduction. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) David Raposo, Google DeepMind and with equal contribution; (2) Sam Ritter, Google DeepMind; (3) Blake Richards, Google DeepMind and McGill University & Mila; (4) Timothy Lillicrap, Google DeepMind; (5) Peter Conway Humphreys, Google DeepMind; (6) Adam Santoro, Google DeepMind and with equal contribution. Authors: Authors: (1) David Raposo, Google DeepMind and with equal contribution; (2) Sam Ritter, Google DeepMind; (3) Blake Richards, Google DeepMind and McGill University & Mila; (4) Timothy Lillicrap, Google DeepMind; (5) Peter Conway Humphreys, Google DeepMind; (6) Adam Santoro, Google DeepMind and with equal contribution. Editor's note: this is part 4 of 5 of a study detailing a way to make transformer-based language models more efficient by dynamically allocating computational resources. Read the rest below. Editor's note: this is part 4 of 5 of a study detailing a way to make transformer-based language models more efficient by dynamically allocating computational resources. Read the rest below. Editor's note: this is part 4 of 5 of a study detailing a way to make transformer-based language models more efficient by dynamically allocating computational resources. Read the rest below. Editor's note: this is part 4 of 5 of a study detailing a way to make transformer-based language models more efficient by dynamically allocating computational resources. Read the rest below. Table of Links Introduction Background Implementing Mixture-of-Depths Transformers


3.1. Defining a compute budget


3.2. Routing around transformer blocks


3.3. Routing schemes


3.4. Routing implementation


3.5. Sampling and 3.6. Training methods Results

4.1. Training, isoFLOP comparisons
4.2. Auto-regressive Evaluation and 4.3. Mixture-of-Depths-and-Experts (MoDE) Discussion and References Introduction Introduction Background Background Implementing Mixture-of-Depths Transformers 3.1. Defining a compute budget


3.2. Routing around transformer blocks


3.3. Routing schemes


3.4. Routing implementation


3.5. Sampling and 3.6. Training methods Implementing Mixture-of-Depths Transformers 3.1. Defining a compute budget 3.2. Routing around transformer blocks 3.3. Routing schemes 3.4. Routing implementation 3.5. Sampling and 3.6. Training methods 3.1. Defining a compute budget 3.1. Defining a compute budget 3.2. Routing around transformer blocks 3.2. Routing around transformer blocks 3.3. Routing schemes 3.3. Routing schemes 3.4. Routing implementation 3.4. Routing implementation 3.5. Sampling and 3.6. Training methods 3.5. Sampling and 3.6. Training methods Results 4.1. Training, isoFLOP comparisons
4.2. Auto-regressive Evaluation and 4.3. Mixture-of-Depths-and-Experts (MoDE) Results 4.1. Training, isoFLOP comparisons 4.2. Auto-regressive Evaluation and 4.3. Mixture-of-Depths-and-Experts (MoDE) 4.1. Training, isoFLOP comparisons 4.2. Auto-regressive Evaluation and 4.3. Mixture-of-Depths-and-Experts (MoDE) Discussion and References Discussion and References 4. Results 4.1. Training, isoFLOP comparisons We first trained models with a relatively small FLOP budget (6e18) to determine optimal hyperparameters (see figure 3). In general, we found that MoD transformers drag the baseline isoFLOP curve "down and to the right". That is, the optimal MoD transformer achieves a lower loss than the optimal baseline, and also has more parameters. A fortunate consequence of this effect is that there exist smaller MoD models that, while they are not themselves isoFLOP optimal for their hyperparameter setting, are nevertheless as- or better-performing than the optimal baseline model while being faster to step. For example, a 220M parameter MoD (figure 3 model #3) variant slightly outperforms the isoFLOP optimal baseline (also 220M, figure 3 model #1), but is upwards of 60% faster to step during training. Crucially, when run on equivalent hardware these two model variants take take approximately the same amount of wall-clock time to train (figure 3). We tested routing every block or every other block, using capacities from 12.5% to 95% of the total sequence. While routing every other block was crucial for strong performance, we found that aggressive capacity reduction was best (gradual improvements were observed when reducing the capacity down to 12.5% of the total sequence, corresponding to 87.5% of tokens routing around blocks, with performance degrading beyond this point). So, it seems the network is robust to significant capacity reductions as long as there is frequent opportunity for full capacity self-attention and MLP computations. Learned routing is crucial, as MoD transformers that use stochastic routing (implemented using a top-𝑘 operation on router weights sampled from a Gaussian distribution) perform drastically worse than both the baseline and normal MoD transformer (figure 3). Depicted in figure 4 is an isoFLOP analysis for 6e18, 2e19, and 1e20 total FLOPs. The trend that FLOP-optimal MoD transformers have more parameters than the baseline continues for these larger FLOP budgets. Notably, there exist MoD variants that are appreciably faster to step than the isoFLOP-optimal baseline (measured as steps per second when training on equivalent hardware) while also achieving a lower loss (in figure 4 we depict normalized FLOPs per forward pass rather than wall-clock step time per se, but from our experiments the two are tightly correlated. A similar plot can be produced showing relative wall-clock step times and the same basic trend is present). Step-wise speed gains come from two sources. First, the FLOP-per-parameter ratio in MoD transformers is less than in the baselines because some proportion of tokens are routed around blocks. So, for a given model size, a transformer requires fewer FLOPs per forward pass. Second, since isoFLOP-optimal MoD transformers are both bigger and achieve a lower loss than the isoFLOP-optimal baseline, there exist smaller MoD variants that perform as well or better than the isoFLOP-optimal baseline, and these variants are faster to step because they are smaller. Altogether, then, there exist MoD transformers that perform as well as isoFLOP-optimal baselines and are faster to step, both because they use fewer FLOPs per parameter and because they use fewer parameters. Figure 4 also reveals another important finding: the optimal MoD transformer is that which uses as many FLOPs per forward pass as the isoFLOP optimal baseline. This finding allows one to directly predict which sized MoD transformer will perform optimally for a given isoFLOP training budget: one just needs to tune the model size for a given MoD configuration (i.e., capacity and routing frequency) to produce a model that uses as many FLOPs per forward pass as the isoFLOP-optimal baseline, and they will have the optimally performing MoD variant for that configuration. Empirically, we find that it is better to add depth than to add width when adding FLOPs to the model. Nevertheless, while the FLOPs per forward pass determines which model will be the isoFLOP optimal, it does not predict whether the optimal loss will improve upon the baseline (see figure 3. Namely, the optimal capacity appears to be empirically determinable. We found that it is best to use 12.5% capacity blocks, every other block. We noticed that MoD transformers had memory savings relative to equivalently sized baseline models at larger sizes, with some variants requiring fewer total devices (i.e., a smaller TPU topology). We did not study this extensively, but we anticipate that as one scales to larger models, these savings could be an important consideration when choosing model variants to train, and could have significant positive effects in regards to the KV cache size during autoregressive sampling. Figure 5 shows the routing decisions for an MoD transformer trained with interleaved routing blocks. Despite aggressive routing around the blocks, transformers are able to achieve performance improvements relative to baselines. We observe patterns that might warrant further study; namely, some tokens appear to engage each block along the transformer’s depth, while others decide to route around blocks whenever possible. Preliminary analyses suggest that the tokens that engage with blocks more frequently are correlated with output predictions that have higher entropy, which possibly corresponds to predictions that are more difficult to make. 4.2. Auto-regressive Evaluation We evaluated MoD variants during auto-regressive sampling (see figure 6). Each model was tested on exactly the same held-out data comprising 256000 sequences (500M tokens). When switching from the top-𝑘 routing method to the predictor-based routing method we observed little performance degradation. As in the training setting, there exist MoD variants that are better performing than the isoFLOP-optimal baseline, while requiring fewer FLOPs per forward pass. These results suggest that the compute savings offered by MoD transformers should translate beyond the training setting. 4.3. Mixture-of-Depths-and-Experts (MoDE) The MoD technique can be naturally integrated with MoE models (together comprising MoDE models) in addition to vanilla transformers. In figure 7 we present results showing that the performance improvments offered by MoD compound with those of MoE. We tried two variants: in staged MoDE, which routes tokens around or towards blocks prior to the self-attention step, and integrated MoDE, which implements MoD routing by integrating “no-op” experts among the conventional MLP experts. The former is advantageous because it allows for tokens to skip the self-attention step, while the latter is advantageous because it simplifies the routing machinery. We noticed that implementing MoDE in the integrated manner was distinctly better than simply reducing the capacity of experts in conventional MoE models, and relying on token dropping to implement residual routing. We believe this is because with the integrated MoDE machinery, tokens explicitly learn to choose the residual path around the experts, as opposed to preferring an expert but being dropped when implemented as a capacity reduction. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

What If AI Could Skip the Boring Parts? Google Researchers Just Made It Happen

Here's the thing about HBO documentaries. The best thing about them is - the production value. The worst thing is - the production value. What's that supposed to mean? Well, you know what they say. After a series of award-winning and jaw-dropping HBO documentaries, such as The Inventor, Leaving Neverland, Going Clear, Beware the Slenderman, and At the Heart of Gold, we were supposed to be blown away with - Money Electric: The Bitcoin Mystery. It didn't happen. Yeah, I've spent more time watching HBO than reading about blockchain. That's why my articles about Bitcoin always play it safe. I share true stories of how Bitcoin impacted the lives of people I know personally or the challenges of Bitcoin acceptance, and similar stuff the MAXIs won't tear me apart for. After spending just one day reading about what blockchain is and how it works from "Cryptocurrency Investing For Dummies" by Kiana Danial, which I couldn't recommend more, I realized that it wasn't rocket science. However, you can't just show a few cubes, which looked as if they were taken out of the movie "Cube," and call it a blockchain day. Back to the production value because it is important. Look, I don't know what happened to the author, Cullen Hoback, after Terms and Conditions May Apply, but at least he didn't stretch Money Electric which was the case with Q: Into the Storm. So, of all people, and after all these years, he was supposed to reveal the biggest Bitcoin Mystery in less than a hundred minutes, give it or take. Who is Miyamoto Musashi? Yeah, I'm fraking (Battlestar Galactica) with you, but for educational purposes only. Miyamoto Musashi was a legendary Japanese samurai "who became renowned through stories of his unique double-bladed swordsmanship and undefeated record in his 62 duels." More importantly, he wrote "The Book of Five Rings," and was "eternalized" as one of the most well-known Joe Rogan's tattoos. By default, the creators of revolutionary things have to be larger-than-life characters themselves. Can you imagine the iPhone without Steve Jobs or SpaceX without Elon Musk? Thus, Satoshi Nakamoto and Bitcoin. There's a scene in the documentary where Cullen Hoback is on the phone with someone, a producer, or whoever is calling the shots, trying to sell his WOW moment—we got him! It was obvious that Adam Back and especially Peter Todd were tossing him around mercilessly. After some time, it stopped being fun and became painful to watch this who-is-Satoshi-Nakamoto game. All that time and money could've been put to much better use in explaining why the block size in blockchain is so damn important. This way we were left with more trivial questions than important answers. Some Mysteries Are Better Left Unsolved Whether Satoshi Nakamoto is one person or a team of people is irrelevant. The myth about Satoshi's 1M BTC is an old and boring story. If you haven't created anything in your life no matter how small and insignificant it may seem to you or enjoyed some magnificent creation first-hand, then you're going to have a problem accepting and realizing why the creations become more important to the creators themselves. Sometimes, I wanted to be published under a pen name or I had to as a ghostwriter. It's both fun and exciting. Gamers care more about their nicks and avatars than anything in the world. I'm no different: Let's wrap this Bitcoin story with the best reply on Reddit. You see, some people have to write one line only and be remembered, while others can write books and be forgotten. And, that's OK. We may never find out who created Bitcoin, but we hopefully know why. And, that's OK too. Till the next HBO documentary and crypto revolutionary, in tech we trust, Bitcoin is a blast! Here's the thing about HBO documentaries. The best thing about them is - the production value. The worst thing is - the production value. What's that supposed to mean? Well, you know what they say. After a series of award-winning and jaw-dropping HBO documentaries, such as The Inventor, Leaving Neverland, Going Clear, Beware the Slenderman, and At the Heart of Gold, we were supposed to be blown away with - Money Electric: The Bitcoin Mystery. It didn't happen. Yeah, I've spent more time watching HBO than reading about blockchain. That's why my articles about Bitcoin always play it safe. I share true stories of how Bitcoin impacted the lives of people I know personally or the challenges of Bitcoin acceptance, and similar stuff the MAXIs won't tear me apart for. After spending just one day reading about what blockchain is and how it works from "Cryptocurrency Investing For Dummies" by Kiana Danial, which I couldn't recommend more, I realized that it wasn't rocket science. However, you can't just show a few cubes, which looked as if they were taken out of the movie "Cube," and call it a blockchain day. Back to the production value because it is important. Look, I don't know what happened to the author, Cullen Hoback, after Terms and Conditions May Apply, but at least he didn't stretch Money Electric which was the case with Q: Into the Storm. So, of all people, and after all these years, he was supposed to reveal the biggest Bitcoin Mystery in less than a hundred minutes, give it or take. Who is Miyamoto Musashi? Yeah, I'm fraking (Battlestar Galactica) with you, but for educational purposes only. Miyamoto Musashi was a legendary Japanese samurai "who became renowned through stories of his unique double-bladed swordsmanship and undefeated record in his 62 duels." More importantly, he wrote "The Book of Five Rings," and was "eternalized" as one of the most well-known Joe Rogan's tattoos. double-bladed swordsmanship and undefeated record By default, the creators of revolutionary things have to be larger-than-life characters themselves. Can you imagine the iPhone without Steve Jobs or SpaceX without Elon Musk? Thus, Satoshi Nakamoto and Bitcoin. There's a scene in the documentary where Cullen Hoback is on the phone with someone, a producer, or whoever is calling the shots, trying to sell his WOW moment—we got him! It was obvious that Adam Back and especially Peter Todd were tossing him around mercilessly. After some time, it stopped being fun and became painful to watch this who-is-Satoshi-Nakamoto game. All that time and money could've been put to much better use in explaining why the block size in blockchain is so damn important. This way we were left with more trivial questions than important answers. Some Mysteries Are Better Left Unsolved Whether Satoshi Nakamoto is one person or a team of people is irrelevant. The myth about Satoshi's 1M BTC is an old and boring story. If you haven't created anything in your life no matter how small and insignificant it may seem to you or enjoyed some magnificent creation first-hand, then you're going to have a problem accepting and realizing why the creations become more important to the creators themselves. Sometimes, I wanted to be published under a pen name or I had to as a ghostwriter. It's both fun and exciting. Gamers care more about their nicks and avatars than anything in the world. I'm no different: Let's wrap this Bitcoin story with the best reply on Reddit . You see, some people have to write one line only and be remembered, while others can write books and be forgotten. And, that's OK. We may never find out who created Bitcoin, but we hopefully know why. And, that's OK too. Reddit Till the next HBO documentary and crypto revolutionary, in tech we trust, Bitcoin is a blast!

I Watched HBO’s ‘Money Electric’ on Bitcoin: Here’s Why You Might Want To Skip It (or Not)

Gledao sam HBO-ov 'Money Electric' o Bitcoinu: Evo zašto biste ga mogli preskočiti (ili ne)

Sledoval jsem „Money Electric“ HBO o bitcoinech: Zde je důvod, proč byste to možná chtěli přeskočit (nebo ne)

Παρακολούθησα το «Money Electric» του HBO στο Bitcoin: Εδώ είναι γιατί μπορεί να θέλετε να το παραλείψετε (ή όχι)

Vi “Money Electric” de HBO sobre Bitcoin: esta es la razón por la que quizás quieras saltártelo (o no)

Ho guardato "Money Electric" della HBO su Bitcoin: ecco perché potresti volerlo saltare (o no)

ビットコインに関する HBO の「Money Electric」を視聴しました。視聴しないほうがよい理由 (またはしないほうがよい理由)

ຂ້ອຍໄດ້ເບິ່ງ 'Money Electric' ຂອງ HBO ໃນ Bitcoin: ນີ້ແມ່ນເຫດຜົນທີ່ເຈົ້າອາດຈະຕ້ອງການຂ້າມມັນ (ຫຼືບໍ່)

Ik heb HBO's 'Money Electric' op Bitcoin bekeken: dit is waarom je het misschien wilt overslaan (of niet)

Я посмотрел сериал HBO «Money Electric» на биткоины: вот почему вы можете захотеть его пропустить (или нет)

මම Bitcoin මත HBO හි 'Money Electric' නැරඹුවෙමි: මෙන්න ඔබට එය මඟ හැරීමට අවශ්‍ය විය හැක්කේ (නැතහොත්)

Men HBO-ning Bitcoin-da "Money Electric" filmini tomosha qildim: Mana nima uchun uni o'tkazib yuborishni xohlashingiz mumkin (yoki yo'q)

Ngibuke i-'Money Electric' ye-HBO ku-Bitcoin: Nakhu Kungani Ungase Ufune Ukuyeqa (noma Cha)

The Preface and Base Structure of a Skip List A skip list is a probabilistic data structure that serves as a dynamic set, offering an alternative to red-black or AVL trees. It supports operations like search, inserts, and delete, which typically have an average time complexity of O(log(n)), although in the worst cases, time complexity will be O(n). Skip lists are memory buffers in various NoSQL solutions, including Redis and RocksDB. The structure of a skip list is composed of multiple layers, referred to as levels. Each level contains roughly half the nodes of the previous level. The first level includes all sorted nodes, and each subsequent level contains about half of the nodes from the level below. This hierarchical setup helps visualize the basic structure of a skip list. As you can see, our skip list consists of the levels. All the values of each skip list are sorted. Also, we must think about the high bound of levels. We take the const 16 because the original paper refers to it, and it ensures log(n) time complexity for lists that contain 2^16 elements. Let’s code the base structure of our skip list. const (
	maxLevel    = 16
	probability = 0.5
)

type node[T any] struct {
	value T
	next  []*node[T]
}

type SkipList[T any] struct {
	compare func(a, b T) int // returns -1 if a is less than b, 0 if a and b are equal, and 1 if a is greater than b
	len     int // number of elements of a skip list
	level int // current level
	head  *node[T] // dummy node
}

func newNode[T any](value T) *node[T] {
	return &node[T]{
		value: value,
		next:  make([]*node[T], maxLevel),
	}
}

func New[T any](compare func(a, b T) int) *SkipList[T] {
	return &SkipList[T]{
		compare: compare,
		level:      1,
		head:       newNode[T](*new(T)),
	}
} This code requires an explanation. Our node has a field next that refers to the node on all the levels. Also, we can denote the function compare that returns -1 if a is less than b, 0 if a is equal to b, and 1 if a is greater than b. We initialize our skip list by 1 level. Search Operation The search operation of the skip list is very simple. We will follow by the highest level before our node is less than the current node; then we down to the next level and so on. We don’t go to the first level. This image illustrates our route. We want to find the node with the value 12. We have to pass through two nodes at level 3, one node at level 2, and two nodes at level 1. In our implementation, we will start from the highest level of our skip list and will go unless the current node  is smaller than the desired node. We will repeat our doing until we get to level 1. func (s *SkipList[T]) Search(value T) (T, bool) {
	current := s.head

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
	}

	if current.next[0] != nil && s.compare(key, current.next[0].value) == 0 {
		return current.next[0].value, true
	}

	return *new(T), false
} Insert Operation The insertion operation is somewhat more complex than the search operation. We introduce a new update variable that records the last node smaller than the inserted value at each level. If the inserted node is already present in the skip list, no action is taken. Next, we need to determine the level to which the new node will be promoted. If this calculated level exceeds the current level, additional levels will be added. We then update the references. The image below illustrates the insertion operation. The node with the value 1 is a dummy node. The current level is 3, but after calling randomLevel, a new level of 4 is determined. The highlighted node is the newly inserted node. The code of the insert operation. func (s *SkipList[T]) Insert(value T) {
	current := s.head
	update := make([]*node[T], maxLevel)

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
		update[i] = current
	}

	if current.next[0] != nil && s.compare(current.next[0].value, value) == 0 {
		return
	}

	level := randomLevel()

	if level > s.level {
		for i := s.level; i < level; i++ {
			update[i] = s.head
		}
		s.level = level
	}

	x := newNode[T](value)

	for i := 0; i < level; i++ {
		x.next[i] = update[i].next[i]
		update[i].next[i] = x
	}

	s.len++
}

func randomLevel() int {
	level := 1
	for rand.Float64() < probability && level < maxLevel {
		level++
	}
	return level
} Delete Operation The delete operation is similar to the insert operation. We use the update variable to store the last node smaller than the node to be deleted. After identifying all relevant nodes, we remove them if the subsequent node is the one to be deleted. If the dummy node points to nil, we will decrement the level. func (s *SkipList[T]) Remove(value T) bool {
	current := s.head
	update := make([]*node[T], maxLevel)

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
		update[i] = current
	}

	if current.next[0] == nil || s.compare(current.next[0].value, value) != 0 {
		return false
	}

	for i := 0; i < s.level; i++ {
		if update[i].next[i] == nil || s.compare(current.next[i].value, value) != 0 {
			break
		}
		update[i].next[i] = update[i].next[i].next[i]
	}

	for s.level > 1 && s.head.next[s.level-1] == nil {
		s.level--
	}

	s.len--

	return true
} In this article, we won’t consider such operations as ceil, floor, etc. As you can see, the implementation is easier than that of balanced trees. However, it requires more memory and cannot guarantee logarithmic time complexity. The Preface and Base Structure of a Skip List A skip list is a probabilistic data structure that serves as a dynamic set, offering an alternative to red-black or AVL trees. It supports operations like search, inserts, and delete, which typically have an average time complexity of O(log(n)), although in the worst cases, time complexity will be O(n). Skip lists are memory buffers in various NoSQL solutions, including Redis and RocksDB. The structure of a skip list is composed of multiple layers, referred to as levels. Each level contains roughly half the nodes of the previous level. The first level includes all sorted nodes, and each subsequent level contains about half of the nodes from the level below. This hierarchical setup helps visualize the basic structure of a skip list. As you can see, our skip list consists of the levels. All the values of each skip list are sorted. Also, we must think about the high bound of levels. We take the const 16 because the original paper refers to it, and it ensures log(n) time complexity for lists that contain 2^16 elements. Let’s code the base structure of our skip list. const (
	maxLevel    = 16
	probability = 0.5
)

type node[T any] struct {
	value T
	next  []*node[T]
}

type SkipList[T any] struct {
	compare func(a, b T) int // returns -1 if a is less than b, 0 if a and b are equal, and 1 if a is greater than b
	len     int // number of elements of a skip list
	level int // current level
	head  *node[T] // dummy node
}

func newNode[T any](value T) *node[T] {
	return &node[T]{
		value: value,
		next:  make([]*node[T], maxLevel),
	}
}

func New[T any](compare func(a, b T) int) *SkipList[T] {
	return &SkipList[T]{
		compare: compare,
		level:      1,
		head:       newNode[T](*new(T)),
	}
} const (
	maxLevel    = 16
	probability = 0.5
)

type node[T any] struct {
	value T
	next  []*node[T]
}

type SkipList[T any] struct {
	compare func(a, b T) int // returns -1 if a is less than b, 0 if a and b are equal, and 1 if a is greater than b
	len     int // number of elements of a skip list
	level int // current level
	head  *node[T] // dummy node
}

func newNode[T any](value T) *node[T] {
	return &node[T]{
		value: value,
		next:  make([]*node[T], maxLevel),
	}
}

func New[T any](compare func(a, b T) int) *SkipList[T] {
	return &SkipList[T]{
		compare: compare,
		level:      1,
		head:       newNode[T](*new(T)),
	}
} This code requires an explanation. Our node has a field next that refers to the node on all the levels. Also, we can denote the function compare that returns -1 if a is less than b, 0 if a is equal to b, and 1 if a is greater than b. We initialize our skip list by 1 level. Search Operation The search operation of the skip list is very simple. We will follow by the highest level before our node is less than the current node; then we down to the next level and so on. We don’t go to the first level. This image illustrates our route. We want to find the node with the value 12. We have to pass through two nodes at level 3, one node at level 2, and two nodes at level 1. In our implementation, we will start from the highest level of our skip list and will go unless the current node  is smaller than the desired node. We will repeat our doing until we get to level 1. func (s *SkipList[T]) Search(value T) (T, bool) {
	current := s.head

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
	}

	if current.next[0] != nil && s.compare(key, current.next[0].value) == 0 {
		return current.next[0].value, true
	}

	return *new(T), false
} func (s *SkipList[T]) Search(value T) (T, bool) {
	current := s.head

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
	}

	if current.next[0] != nil && s.compare(key, current.next[0].value) == 0 {
		return current.next[0].value, true
	}

	return *new(T), false
} Insert Operation The insertion operation is somewhat more complex than the search operation. We introduce a new update variable that records the last node smaller than the inserted value at each level. If the inserted node is already present in the skip list, no action is taken. Next, we need to determine the level to which the new node will be promoted. If this calculated level exceeds the current level, additional levels will be added. We then update the references. The image below illustrates the insertion operation. The node with the value 1 is a dummy node. The current level is 3, but after calling randomLevel, a new level of 4 is determined. The highlighted node is the newly inserted node. The code of the insert operation. func (s *SkipList[T]) Insert(value T) {
	current := s.head
	update := make([]*node[T], maxLevel)

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
		update[i] = current
	}

	if current.next[0] != nil && s.compare(current.next[0].value, value) == 0 {
		return
	}

	level := randomLevel()

	if level > s.level {
		for i := s.level; i < level; i++ {
			update[i] = s.head
		}
		s.level = level
	}

	x := newNode[T](value)

	for i := 0; i < level; i++ {
		x.next[i] = update[i].next[i]
		update[i].next[i] = x
	}

	s.len++
}

func randomLevel() int {
	level := 1
	for rand.Float64() < probability && level < maxLevel {
		level++
	}
	return level
} func (s *SkipList[T]) Insert(value T) {
	current := s.head
	update := make([]*node[T], maxLevel)

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
		update[i] = current
	}

	if current.next[0] != nil && s.compare(current.next[0].value, value) == 0 {
		return
	}

	level := randomLevel()

	if level > s.level {
		for i := s.level; i < level; i++ {
			update[i] = s.head
		}
		s.level = level
	}

	x := newNode[T](value)

	for i := 0; i < level; i++ {
		x.next[i] = update[i].next[i]
		update[i].next[i] = x
	}

	s.len++
}

func randomLevel() int {
	level := 1
	for rand.Float64() < probability && level < maxLevel {
		level++
	}
	return level
} Delete Operation The delete operation is similar to the insert operation. We use the update variable to store the last node smaller than the node to be deleted. After identifying all relevant nodes, we remove them if the subsequent node is the one to be deleted. If the dummy node points to nil, we will decrement the level. func (s *SkipList[T]) Remove(value T) bool {
	current := s.head
	update := make([]*node[T], maxLevel)

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
		update[i] = current
	}

	if current.next[0] == nil || s.compare(current.next[0].value, value) != 0 {
		return false
	}

	for i := 0; i < s.level; i++ {
		if update[i].next[i] == nil || s.compare(current.next[i].value, value) != 0 {
			break
		}
		update[i].next[i] = update[i].next[i].next[i]
	}

	for s.level > 1 && s.head.next[s.level-1] == nil {
		s.level--
	}

	s.len--

	return true
} func (s *SkipList[T]) Remove(value T) bool {
	current := s.head
	update := make([]*node[T], maxLevel)

	for i := s.level - 1; i >= 0; i-- {
		for current.next[i] != nil && s.compare(current.next[i].value, value) < 0 {
			current = current.next[i]
		}
		update[i] = current
	}

	if current.next[0] == nil || s.compare(current.next[0].value, value) != 0 {
		return false
	}

	for i := 0; i < s.level; i++ {
		if update[i].next[i] == nil || s.compare(current.next[i].value, value) != 0 {
			break
		}
		update[i].next[i] = update[i].next[i].next[i]
	}

	for s.level > 1 && s.head.next[s.level-1] == nil {
		s.level--
	}

	s.len--

	return true
} In this article, we won’t consider such operations as ceil, floor, etc. As you can see, the implementation is easier than that of balanced trees. However, it requires more memory and cannot guarantee logarithmic time complexity.

Skip List From Scratch: A Guide

The Crucial First Step Toward AI Integration Companies Can’t Skip

Crypto isn’t going anywhere and so is the case with crypto launchpads. Why is it that launchpads offer that developers find valuable? Launchpads offer a variety of projects, tools, and resources that help developers take the project from the ground. The obvious question would be what are the top launchpads that you should look out for in Q3 2022 and beyond? This article will discuss the best in the game and why these and not others. What are Crypto Launchpads? For those who don’t know what a launchpad is. Launchpads are platforms that allow crypto developers to get investment before the tokens are publicly released. On the other hand, investors can also find out new interesting projects to invest in. In simple terms, you can consider them as a marketplace where investors and crypto project developers can connect. But why do we need launchpads when you can directly connect with investors? Well, it turns out that before the advent of launchpads in 2017, crypto projects used to hold private sales on their website. This process was called and investors had to hold the token before it is publicly released. Initial Coin Offering (ICO) However, most of the projects turned out to be scams, costing a huge loss and trust issues among investors. To cope with the underlying problem, crypto centralized exchanges took measurements and take on the job of researching before partnering with any project. This approach worked since the exchanges already had a user base and investors to invest in projects. Meanwhile, decentralized exchanges emerged offering opportunities for investors to buy in new projects. This token sales process is called the initial DEX offering, and the platform or service that DEXes uses to provide this is also called a launchpad. In simple terms, any platform that raises funds for crypto projects by allowing you to buy tokens before they are publicly released is called a Crypto launchpad. Launchpads offer one very clear advantage to developers and crypto investors trying to break into the market. They offer the best insights and data on new and upcoming projects to participate in their presales which typically means less investment. Let’s dive right into it for you to choose your next crypto launchpad 1- Binance Launchpad Launched in 2017, Binance didn’t hold itself back from being more than just a cryptocurrency exchange. Since 2017, Binance launchpad has come a long way and is widely adopted by the community for high contributions in the DeFi and crypto industry. It also has it's own native token. Why ? Binance is unique in its approach as it gives users first-look access to new crypto projects along with awareness around the projects to help crypto enthusiasts in decision-making. Projects like , , , and have successfully raised millions of dollars through public sales on the Binance platform. Binance launchpad Reserve (RSR) BlaBlaGame (BLA) Serenity(SET) Bet2ken(BTK) 2- Seedify. Fund Are you looking for a crypto launchpad but in the gaming world? No worries. is a leading Blockchain gaming launchpad and it has already launched 50 projects successfully with the highest average ROI in the industry. Seedify It helps you invest in Blockchain games, NFTs, and Metaverse like 1,2, 3. Yes, it is that simple. How does that happen? Access to leading projects Keen due diligence of the projects It helps you purchase tokens and NFTs early Offers staking and farming as passive income Upcoming IGO and INO on the platform are ARtonar, Haunted Space, Realm Hunter, Aradena Battlegrounds, etc. 3 - Kaizen Launch Booster is a platform for token offerings and the launch of crypto projects but what makes it different from the rest? It just isn’t a platform for tokens and the launch of projects but it takes it to the next step and provides the necessary technical tools to attract investment with the establishment of an ecosystem that is filled with seasoned investors. Kaizen Launch Booster Three major highlights of the platform: Locked token management: Ability to distribute locked-token right after purchasing. Double scoring: A streamlined audit process before listing any project. Well-versed and loyal community 4 - Polkastarter Launchpad One of the best crypto launchpads to make it to the list. Launched in 2020 and became an instant hit with over 200,000 enthusiastic investors. The distinctive features of the include: Polkastarter launchpad Top-notch auction system Private pools that are password protected Seamless whitelisting Easy token swaps Regular price alerts Other than these amazing features, it helps new projects in raising capital and building dedicated communities. Up until now, 108 projects have been launched by Polkastarter amounting to $48,475,044. 5- Trustswap makes it to the list because of its immense support for users. It offers a lucrative offer to the users. You can stake a minimum of 4000 SWAP tokens (the native token of trust swap) to participate in sales of upcoming projects. Trustswap Projects selected by Trustswap have numerous benefits including but not limited to large-scale publicity, community trust, support and visibility, and market support. Trustswap raise around $20 million for cool emerging projects like Chain Games, Glitch Finance, SOTA, and BitCashPay since its launch. Conclusion To conclude, Crypto is here to stay and so are these widely popular and populated crypto launchpads. The choice boils down to your project type and which launchpad will be the best fit for your project’s launch. The list helps you make an informed decision with all the pros and cons listed above. Disclosure: This story was submitted to HackerNoon by an independent contributor. Hence the information contained therein has also been researched and compiled independently.

5 Crypto Launchpads Not to Skip in Q3 2022

In the nascent stages of and its accompanying digital assets, smart contracts were the key to unlocking the vast potential of this newfangled innovation. Smart contract is termed as an agreement between the buyer and seller which consists of a set of and Whenever a user requests a transaction, these functions and data get executed automatically. No third party is needed to mediate or verify the transaction. This makes for a more secure, trustworthy, and efficient system overall. blockchain technology functions data. Features of a Smart Contract Audit When it comes to smart contracts, trust is of the utmost importance. Because there is no third party involved in verifying or enforcing the contract, it's essential that all parties involved can trust that the code will do what it's supposed to. This is where a good smart contract auditor comes in. The auditor reviews your code for any security vulnerabilities or potential exploits that could lead to a loss of funds or any smart contract compromise. The auditor also checks for any coding mistakes that could cause the smart contract to malfunction. By identifying and fixing these issues before anything goes wrong, you can rest assured that your smart contracts will run smoothly and without any breach incident. Due to their self-executing nature, smart contracts are often deployed without a second thought given to the security of the code. This can lead to disastrous consequences. For this reason, it is imperative that businesses and individuals who wish to utilize smart contracts conduct a thorough audit of the code before deployment. What is Smart Contract Auditing? The process of examining and verifying the functioning and security of a smart contract is known as . The auditor looks for any vulnerabilities that could be exploited and cause losses to the contracting parties. They also check for compliance with industry best practices and regulatory requirements. smart contract auditing Importance to Firms Given the security nature of smart contracts, firms are increasingly looking to them as a way to streamline their operations and reduce costs. Smart contracts can automate many processes that would traditionally require human interaction. This not only speeds up transactions but also eliminates the possibility of human error. In addition, because smart contracts are immutable, they provide a high degree of certainty when it comes to enforcing contractual agreements. For example, when the injured person's rights or interests are adequately safeguarded and he or she is not disadvantaged, a contract can be signed without the necessity for costly legal fights. Who Is It For? Smart contract auditing is not just for firms, though. Individuals and other organizations can also benefit from the security and efficiency that this process brings. In fact, many people see smart contract auditing as a key piece of the overall blockchain puzzle. It is an essential step in ensuring that this new technology lives up to its potential and revolutionizes how business is conducted worldwide. Firms That Conduct Smart Contract Auditing There are several firms that offer smart contract auditing services. Some of these include: Astra Security Chainalysis Secureblock Solidity Finance BitGo ChainSecurity ConsenSys Certik Defisecurity Monax OpenZeppelin IBM KPMG Hosho Group QuantStamp Each of these companies has extensive experience in the field and employs teams of experts who can help you assess and verify the security of your . smart contracts How to Choose a Smart Contract Auditor? When looking for a firm to conduct your smart contract auditing, it is important to consider the features that they offer. Some of the key features to look for include: Experience Technology Services Expertise Security Compliance Experience and expertise are essential, as they ensure that the auditor has the knowledge and skills necessary to find any potential and also to ensure that your contracts are safe and compliant with industry practices. This way you can be confident that your contracts will be handled in a professional and secure manner. vulnerabilities in your code Technology is also important, as it allows auditors to use cutting-edge tools and techniques to assess smart contracts. Finally, services offered should include a wide range of options, such as consulting, training, and support. This ensures that you get all the help you need in order to make the most informed decision about your smart contract security. Pros And Cons Of Smart Contract Auditing Like any other process, smart contract auditing has its pros and cons. Let's take a look at some of the key advantages and disadvantages of this important step in : blockchain development Advantages: Smart contracts are more secure and efficient than traditional contracts. Smart contract auditing can help reduce or eliminate the need for costly legal disputes. Auditors use cutting-edge technology to assess contracts for security vulnerabilities. Disadvantages: Smart contract auditing is an essential but time-consuming process, which can delay transactions. Not all firms offer a wide range of services, so it is important to do your research before choosing one. Summing Up Smart contract auditing is a key process for ensuring the security and efficiency of your blockchain system. It's crucial for companies wanting to cut costs and speed up operations. Individuals and other organizations can also benefit from smart contract auditing. When looking for a firm to conduct your audit, be sure to consider the features that they offer. These include experience, expertise, and technology. Firms that offer a wide range of services are best suited to meet your needs.

Why You Should Never Skip a Smart Contract Audit

As a leader of an organisation, the question “what needs attention” was always top of my mind. I realised that without taking real inputs from people in the organisation, any decisions I make and the direction I provide will be nothing but my own biased view of what people want, not what they need. Looking down from 10000 ft above the ground, you may see the greenery of the forests, ice-covered mountains, beautiful terrains, and vastness of the oceans. It’s only when you get to the ground, the rough patches start to surface. Once you notice the new reality, you can shift your perspective to align with the real world. I realised that this is so true at work too. If I let hierarchy magnify the distance that I have from my people especially those who do not report to me, I can never learn about the issues that impact people on a day-to-day basis. Are they looking for more empowerment, recognition, engagement, growth or are they happy with the way things are? How are they learning? Without diving deep into people’s minds, I can never develop the intelligence and perspective required to stay close to reality. So I decided to go to the source, hear it from their own mouth without any presumptions. Before I share my learnings from skip-level meetings, you may be wondering why I could not trust my direct reports to provide me this information. What’s truly so valuable about skip-level meetings that I had to skip my direct reports and speak to the people reporting to them. Put another way, why was it worth my time? What value do skip-level meetings provide? If I wanted to create a map for my organisation that can help people determine which paths to avoid and which ones to take to be successful in the long run, I had two options: Option 1 - Design all possible paths and subject them to my interpretation based on my beliefs and assumptions. Aside from being highly complex, it will be limited to my circle of competence and those of my direct reports Option 2 - Draw on the collective knowledge of all the people in my org and overlay their knowledge with my experience to design this map Aside from the fact that option 2 helps unearth real problems, it’s very effective in finding patterns that either enable people to move forward or hold them back. If I could access the personal map every person uses to navigate the organisation, I could design a strategy that enables everyone to be more effective in their work. As Peter Drucker said in his book, The Effective Executive The more aware we are of our basic paradigms, maps, or assumptions, and the extent to which we have been influenced by our experience, the more we can take responsibility for those paradigms, examine them, test them against reality, listen to others and be open to their perceptions, thereby getting a larger picture and a far more objective view In any organisation, all strategies and tactics are useless without clarity on what we are trying to achieve and why it’s the right direction for the organisation and its people. Accessing this map is fundamental to uncover hidden treasures and design a long term strategy that’s rooted in the collective power of minds. Each individual MAP gave me access to information that could be put to use to draw patterns of behaviour and thinking in the organisation. Notice the idea of meeting my indirect reports is not to spy on people who report to me. Though done right, it can also be used to draw valuable insights about how they feel about their manager. Make an attempt to inquire about their manager directly and they will never share true feedback. While learning about their M-mindset, A-awareness and P-priorities (their personal map) helps get below the surface to uncover reality. Another positive side-effect of meeting them face-to-face that I didn’t realise at the time was the connection and the trust that it established. Knowing each person by name, learning about what they value, and how they are doing helped me establish a deeper bond with each one of them. I truly found value in meeting every single one of them and looked forward to each such interaction. MAP framework for effective skip-level meetings Asking the right questions and giving an opportunity to the other person to speak is at the heart of every useful conversation. L. David Marquet suggests in his book, Turn The Ship Around If you walk about your organisation talking to people, I’d suggest that you be as curious as possible. As with a good dinner table conversationalist, one question should naturally lead to another. The time to be questioning or even critical is after trust has been established. With that in mind, let’s understand what this framework provides: M - Mindset Mindset reflects their feelings. These feelings provide indirect answers to a lot of questions that are otherwise very difficult to analyse Do they feel fearful or safe to take risks Are they empowered to make decisions or seek approval for every action Do they get the required support that helps them move forward or find themselves struggling every step of the way Do they feel recognised for their work or ignored without appreciation How do they feel being part of their team and this organisation Do they find the culture of the company conducive to learning or restrictive in growth How do they perceive fairness in the organisation Once you understand how they feel without judgement or opinion, you can connect their mindset to levels of awareness, which takes us to the next step in the map. A - Awareness Awareness reflects their understanding of the organisation and their relationship to it. It offers key insights into the awareness level of the person Are they aware of their personal goals Do they understand their goals as a team and how are they relevant Can they find a connection of their team to the vision and mission of the org How do they view the success of the company and their personal success Do they see how their work is aligned with the direction of the company What do they feel about other teams and functions Do they find value in engaging outside their teams How can they contribute beyond their assigned responsibilities Understanding this connection of an individual to the larger purpose of the organisation opens the gateway to finding their priorities. P - Priorities Priorities of a person reflect what they truly care about. Every person has different core needs at work that shapes how they perceive information and experience the world around them. It’s how they behave, decide, and act in the workplace. Learning about their priorities requires understanding What’s top of their mind What do they find valuable in their work How do they react to new information or changes around them At this moment, do they need growth, stability or flexibility This map serves as a mental model to help connect the dots to what’s working for people and what prevents them from doing their best work. Now that I had a framework in place to derive value from skip-level meetings, I needed to put it into practice. I had to determine the right frequency that will add value to our conversations. How often do you need skip-level meetings? If you are a director, VP, or someone in a CXO position, you may have a few direct reports (other managers) with a large number of people reporting to them. Depending on the number of indirect reports, balance out the schedule such that it’s not too frequent nor too far apart. I preferred meeting everyone once a quarter for 30 mins. You can do it for 45-60 mins as well depending on your comfort level and change it as you go once the momentum is established. Assume you have 80 indirect reports and you choose to meet them once a quarter for 30-45 mins, all that’s needed is 5 hours each week on your calendar, which is nothing compared to the value it provides in the long run. With framework and frequency determined, the next step is to start on the right note. How to start skip-level meetings? There are 2 key elements of putting your intent to action: 1 - Messaging it right If the process of doing skip-level meetings is new to your organisation, it’s extremely important to inform your managers about your intent to meet their direct reports and explain what it means to you and them. Sharing your perspective on how it helps the organisation sets the right tone for the purpose of the discussion. Keeping them in the loop on this decision not only builds trust, it avoids the confusion and anxiety that might come from not knowing. 2 - Scheduling it It doesn’t matter which person you pick to start, but it’s important to schedule it for every single member of your org. The information that you are meeting indirect reports will spread faster than you can imagine. Try to do it across teams picking people from different roles and backgrounds every week to avoid people feeling left out or biased. The invite should be thoughtful and precise with the objective of the meeting as an intent to collaborate instead of a one-sided conversation. From my experience, a recurring invite from the beginning also establishes your seriousness and commitment to do it right and not make people consider it as a one-off thing. How to make the most of skip-level meetings? I observed myself going through 3 stages of skip-level meetings. The time in each stage varied from person-to-person depending on how they viewed these meetings and the effort they were willing to put in to add value to the conversation. To be honest, not every interaction was super useful and some were quite draining due to the drama in the conversation. However, people willing to engage in a meaningful dialogue provided value that far outweighed some little episodes of emotional stress along the way. To make the most of skip-level meetings, it’s important to observe how your relationship with each person goes through these 3 stages. As skip-level meetings are far apart, typically a few months, the context from the previous meeting brings in a ton of value to the current discussion. I preferred taking notes along the way to help me remember key points from the discussion for our next meeting. It’s also very interesting to notice how their map (mindset, awareness, priorities) has shifted in the last few months trending in the positive or negative direction. Stage of curiosity Each person has a different personality. Some are quiet, others loud and expressive. Some struggle with words, while others can impress you with it. Some are honest and straightforward, while others will try to exaggerate or blow a situation out of proportion. Some enjoy their work without the need for attention while others do work to get the necessary attention. Understanding each person in terms of their personality is an important element to make the most out of these discussions and form a balanced view of their thoughts and opinions. Always start with yourself. Share a bit about your passions, what motivates you, your background, what brought you to this company, and any other personal details that might be relevant for the person to feel comfortable. To establish the connection and learn about each person’s personality, actively listen by going beyond words to their body language, tone, and intent. These non-verbal cues will provide invaluable information to understand each person better. Some good questions to ask in this stage: How would you describe yourself I am curious to know about your past work, family and education What’s most valuable about work What do you hate or find extremely frustrating about your job Who is the most meaningful person in your life and why Who do you enjoy working with What does learning mean to you What’s been your biggest struggle so far and what did you learn from it How do you find traveling to work everyday With a peek into their world, it’s the right time to learn more about their maps. Stage of learning Getting people to open up is a really tricky job. It will require you to play a little bit of a detective and push the right levers to get to the underlying message. Be flexible in your thinking without being biased or quick to jump to conclusions about what you hear. Take time to process the information and try to infer patterns from these discussions. Key is to ask some great follow-up questions to some of these initial questions: How do you feel about your work - positive, negative or neutral How do you see yourself contributing to decisions in the team What do you like about your role and the organisation Do you see yourself as part of a team What are some issues that bothered you recently What’s the biggest value you bring to your team How do you see the connection of your work to the goals of your team and the mission of your org What’s the most important thing in your life right now that you value and want to prioritise. What are you willing to trade-off to make it happen How do you invest in learning? Do you find company’s environment supportive in enabling it What’s the biggest mistake you made recently and what did you learn from it? How did you feel in the moment - criticised, supportive, fearful, empowered One person you look up to beyond your team - from other teams and functions While there are hundreds of questions on this theme, I guess you get the idea of the kind of questions that can get you started. Don’t scratch the surface, try to go as deep as possible without pushing too hard. Know when it’s time to back off. Stage of trust After multiple conversations, when people see that you truly care about their well being, there’s a shift from answering questions to seeking answers, openness to engage and share without apprehension. As trust starts to emerge and you sense a level of comfort with your indirect report, it’s time to seek some honest feedback. Ask about how you can do better, what should you do more, start doing or stop doing to enable them to do their work better. They may have nothing to say the first time around. Reframe it, ask again and be silent until you get them to speak up. Silence is a very powerful tool at your disposal. Use it as much as possible to give a chance to the other person to express. Don’t get defensive or try to explain if you get some critical feedback, simply hear them out and chew on the information later. Skip-level meetings done right can be a powerhouse of useful information, a chance to connect and care, learn and grow, and do the right thing for your organisation and its people. It will be the best investment of your time as a leader. Do you see a value in skip-level meetings? What are your learnings from them? Previously published at https://www.techtello.com/skip-level-meetings/

Skip-Level Meetings And How To Run Them

Setting up a skip level meeting for interns can be a great way to add value to the intern, their mentor, and the mentor’s manager. Internships are great for a multitude of reasons. They provide valuable career experience to the intern, they can be a great recruiting tool for the company, and they provide helpful growth opportunities to the employees. A skip level meeting between the intern and their mentor’s manager helps you take advantage of those growth opportunities. Assuming that you are the manager, a skip level meeting allows you to: Practice having skip level meetings Build a relationship between yourself and the intern Provide feedback to the mentor Get feedback for yourself, the team and the company Enhance the intern’s experience Here’s a template for running a skip level with an intern. We’ll assume a summer internship and plan for 3 meetings. Discuss skip levels with the mentor If you’re unfamiliar with skip levels this gives a good overview. The first rule of skip levels: do NOT schedule a skip level without first discussing it with the mentor. Ever. If done poorly it can undermine the mentor and create poor team dynamics. Execution matters article . . Successful skip levels are built on trust and transparency Start off on the right foot by discussing it with the mentor. Tell them you’d like to set up a skip level Discuss your motivation and intention, let them know you want to support them and the intern Ask them what they think, and if they have any questions or concerns Involve them in choosing the topics and questions for the meetings This should feel like a collaboration between you and the mentor. Meeting invite and first meeting questions The first meeting is the intro meeting. Here are some good topics and questions for a first meeting: What a skip level is and what your role is Reinforce the mentor’s primary role and your secondary role Talk about transparency and mutual support Build rapport and get to know each other Tell me a little about yourself? What do you like to do for fun? What attracted you to our company? Don’t forget to go both ways and open up about yourself Answer any questions about the company, team, or product Discuss what the intern’s goals are for the internship What are they hoping to get out of it? What specific skills are they hoping to improve? Here’s a sample meeting invite: Hi, Putting some time on the calendar for us to chat. This is an opportunity for you to ask me any questions about your internship, the team, or the company. It’s also a good time for me to get feedback from you on how it’s going so far. I’ll prepare a few questions for you. For our first meeting, I’d like you to think about what the #1 thing is that you’d like to get out of this internship. Please let me or your mentor know if you have any questions about this meeting. Cheers, Fabian p.s. The other interns will get a similar invite and your mentor is aware that we’re having this meeting. My goal is to meet with you 2-3 times during your internship. Second meeting The second meeting is the most flexible. They’re in the middle of their internship and hopefully hitting their groove. They should be making progress on their project and feeling more comfortable. Here are some possible topics for the second meeting: A career conversation What does your ideal job look like? What skills do you think you’d need for it? What do you dream of doing? What questions do you have about what it’s like to do this full time? Answer any questions they have What would you like to know more about? What parts of our team or business are you most curious about Get some feedback on how they and their mentor are doing What’s your process for getting unstuck? What has been your experience when you’ve had questions for your mentor or others on the team? What has been most surprising to you so far? What’s been the most fun so far? What’s been less fun? Third and last meeting The third meeting is the wrap-up. This is probably your last chance to get feedback and to provide an exceptional experience. Hopefully everything went swimmingly. They did a fantastic job, you’re excited to make them a full-time offer, and they’re excited to accept! Of course, that’s not going to happen all the time. Even if the picture isn’t as rosy you still want them to have the best experience possible. Whether it’s an intern, an interview candidate, or an office visitor, treating them well says a lot about your culture. Here are some topics for the last meeting: Reflect on the experience How do you think you did? What do you think you did well? What do you think you could have done better? How did we do on your goals for the internship? What are you most proud of? Get feedback for yourself, the mentor, and the organization What’s one thing you think your mentor did really well? What’s one thing you wish they had done differently? Same question for me. What would you do to improve our internship program? What’s one thing you think we should do differently as a company? Appreciation and continued support Thank them for their contributions. Be specific! Wish them luck in the next step in their career (e.g. finishing school, next internship, etc.) Offer them continued support if they have any further questions down the road. Final thoughts Internships offer learning and growth opportunities for everyone involved, not just the intern! A skip level meeting is a great tool to enhance the experience. What are some of your favorite skip level questions for interns? me and let me know. Contact Previously published at https://managingdev.com/skip-level-meetings-interns/

How To Run Great Skip-Level Meetings   For Interns

If you've been on LinkedIn anytime in the past several months, you've probably come across the infamous "certification post." Users show off their certificates or "statement of participation" in almost anything, from LinkedIn Learning courses to attending webinars or completing a coding quiz. I'll admit, I'm guilty of this as well, and I'll explain why below. Understanding the Phenomenon Certificate posts have several powerful drivers: The bandwagon effect / Fear of Missing Out (FOMO) Social media algorithms Engagement-driven business models First off, one post can easily garner tens of thousands of views, and this validation sparks a primal "fear of missing out" in viewers. What if everyone has a certificate except for me? And so, the next person gets a certificate just to share it on social media and feel like they're part of the crowd. Secondly, most social media algorithms include a "positive spiral" of engagement. Feed optimization algorithms include inputs like "number of likes" and "number of comments," so that popular posts get moved to the top, such that the psychological FOMO of certificate posts is algorithmically fueled as well. Finally, social media sites like LinkedIn want growth and engagement at any cost. Without non-stop growth, social media platforms join the endless graveyard of failed experiments. As I've shown, certificate posts simply "work," in terms of getting engagement. This is part of why LinkedIn acquired Lynda for a whopping . one-and-a-half billion dollars LinkedIn Learning certificates now fuel massive engagement, with a major caveat: This engagement isn't by decision-makers, but rather by other students. What this means is that, while you're likely to get thousands of views, you won't actually get the attention (and more importantly, admiration) of those in a position to hire you. The Alternative For every "certificate post," there's another post complaining about certificate posts. Lest this article succumb to the same irony, I'll present an alternative. If we look at , we'll see that there's still massive growth in global advanced analytics employment. data science career trends in 2020 In other words, there's a lot of opportunity for those with the , not those with the certificates. skills This growth isn't over, as there are still tens of thousands of unfilled data science roles. To be one of the lucky few who lands such a role, you need to share content that shows you're both passionate and skilled, which certificates typically fail to do. First, pick a topic you're interested in. Then, analyze the associated data, creating insightful visuals and commentary (basically, do data science). Finally, share these insights. We all know that shared articles don't get read by most viewers (except for the headline), so sharing a dashboard instead is a fantastic way to show your passion and skill in a format that attracts attention. For inspiration, you can look to the dashboard mentioned above, and other analyses like or . data science career trends diversity in big tech coronavirus search trends

Dear Aspiring Data Scientists: Skip the Certificates, Do This Instead

November 7th 2019 Don’t you just CSS? HATE Many developers seem to go through this phase with CSS because it’s not as appealing and fun as learning a programming language. I did too for sometime, especially when I was first learning it, but as I began to understand it, I realized it wasn’t as bad as I had initially thought (or I’m numb to the pain by now). In any case, I get it, so do millions of other developers out there, but in reality, love it or hate it , and understanding it can make all the difference in your life as a web developer. So here are some reasons why shouldn't skip learning CSS. CSS is here to stay You only need to learn it once Like any other skill, once you know CSS, you know CSS, and the best part is that no one can take that away from you. Think of it like learning the alphabet, it seems irrelevant when you are first learning it, but quickly you come to see that everything you do subsequently will ultimately depend on knowing it. CSS is a fundamental pillar of the web, and like any fundamental thing, a deep understanding lays a foundation that will last a lifetime, assuming it will always stick around. Get it in now, and know it forever. Bootstrap is built with CSS in Bootstrap, or any other CSS framework, is completely built using CSS. Everything This may not matter to some, but the truth is that if you’re a developer, you are very likely building web app components on a regular basis. Understanding how the frameworks that you use work and how they are built, will help your workflow tremendously when you are actually using them. The key point here is that understanding the underlying CSS will help you build with more ease with Bootstrap because you will understand the underlying principles that it is built with. Your value as a developer This is a longer point, but it is by far the most important. I will summarize it with some backstory. I periodically visit San Francisco, and as you may guess, I have met a lot of web developers during my visits. One in particular that comes to mind, a lead Netflix developer. I met this person sharing an Lyft of all things, and we began to talk about his work. Eventually, I ended up asking him what skills he believes are the most important for a fresh web developer to have. he recommended CSS, mentioning that he has seen many devs with rigorous engineering backgrounds, but he has seen few that actually know or care to know CSS. Without hesitation Mind-blowing right? Many other developers I have spoken with have said this as well. One thing most devs I’ve met have in common is that they all HATE CSS but recommend learning CSS for anyone coming into the industry. This is not just one or two instances, but literally all of them have confessed in one form or another that CSS is the bottleneck  for many devs, simply because no one wants to take the time to learn it. This should not come  as a surprise to anyone who has spent any time using CSS, it is not fun to use or learn. This is not because there is no intrinsic value in learning it, but because CSS is a completely different language when compared to ‘normal’ programming languages. The way things are done in CSS are vastly different than the way things are done in your standard programming languages. Programming is logic codified whereas, CSS is ...not. With CSS, logic will not help you solve a problem as much as knowing certain ‘tricks’ will, and that IS the frustration of learning and using CSS, and ultimately, that is the reason for the reluctance of many to learn and use it. Now, with all of this being said, why should you learn CSS? Because most people learning it! Simple as that. HATE Your advantage is that you can learn a skill that many are often reluctant to learn, which will boost your value as a developer. Simply put, CSS skills are fairly rare, and anyone that possesses them standout. will Although, I am constantly in a learning phase as a web developer, the need for CSS can’t be ignored. Like any skill, once you have it, it is yours to keep, and knowing it will help you tremendously when it comes time to branch out to less fundamental tools. After all, the web is designed with CSS, and knowing how to design the web on a deep level will only increase your developer value. Would I prefer the web to entirely GUI-less and terminal based? Yes. Will that happen? No. So as far as I can see, the most important thing you can do as a dev is learn CSS.

Stories, News & Mentions about SKIP

Categories

Trending Topics

Stories, News & Mentions about SKIP

Skip’s Stories on HackerNoon

Skip's News & Mentions

Categories

Trending Topics