Preset

preset

Hello My organisation had a directory structure for projects like: grunt/Gruntfile.js grunt/package.json grunt/node_modules/ web/index.html (well, .php) web/js/script.js Which had been fine with our for SCSS, JS etc. grunt toolchain When I added JSX and via I found I encountered an react browserify “Error running grunt-browserify” >> Error: Cannot find module ‘react’ from ‘/directory/of/source/jsx’ There may be a simple way of telling (or similar, I had the same errors when investigating alternatives) the location of the node_modules/ for modules, presets and plugins, but until I find out what it is, I have reorganised to: grunt-browserify /Gruntfile.js /package.json /node_modules web/index.html web/js/react-app.js And changed my to reflect that the web/* source files folder live in a child directory (web/) rather than a sibling (../web/). Gruntfile.js >> Bundle web/js/react-app.js created. I would prefer to have not moved my directories around, but for now, my OCD can live with a bit of pollution of the parent directory.

Reorganising a directory structure to solve “Cannot find module/preset/plugin” in grunt with babel…

Blockchain-based casinos can’t stop making headlines, аs they bring anonymity and (what players love more than anonymity!) almost instant payouts. Now that “AI” is the word for pretty much everything, it’s not a coincidence that the iGaming industry is among the first adopters. Accroding to Casinos Blockchain, the online gambling market will reach approximately $245.4 billion by the end of the calendar year, and AI is very likely to be “guilty” for a significant portion of that growth. $245.4 billion very likely What Makes AI Gambling Different? For many obvious reasons, you may wonder, ‘What’s common between casinos and artificial intelligence?’. Well, casinos nowadays use AI to analyze player habits – play frequency, game preferences, and even betting speed. Algorithms later recommend titles or offer bonuses made to “check” specific preferences, rather than shoot in the dark and “nailing the sweet spot” every now and then. To make it look simple, we can say that game recommendations function pretty much like Netflix suggestions – but for gambling. Those who prefer roulette could receive roulette-specific bonuses rather than the standard promotions (which they may not be interested in, especially if they don’t play slots). The advanced targeting can benefit both parties, but let’s be honest – it wouldn’t be applied if the outcomes weren’t financially backed. And as for the result – 5-15% boost in revenue for crypto casinos – not bad, right? 5-15% boost It is a win-win scenario – casinos boost loyalty while players enjoy their sessions much more frequently. The Safety Net While AI can definitely contribute to better gameplay, it can also power an “active handbreak” – responsible gambling tools that can actively watch player behavior and spot problematic patterns. Systems can flag unusual activity, such as sudden betting increases from typically more “modest” punters. The latter can receive notifications when approaching deposit limits, which are helpful reminders about pre-set risk thresholds. The interventions balance enjoyment and protection, letting players maintain awareness without spoiling the fun. Security Gets Smarter Blockchain already brings transparency to gambling. AI gets it even further. Machine learning models can predict game outcomes, betting frequencies, and login locations to spot cheating or collusion right away. Security systems flag logins from new devices or unusual IP addresses. The result? A safer gambling environment. The combination works in a full symbiosis – blockchain verifies every transaction while AI warrants system integrity. The “dynamic duo” makes players feel safer keeping funds on these platforms. Support Without The Wait Endless support queues can now be avoided, as AI chatbots now handle simple queries instantly and are available 24/7. For more complicated issues, bots transfer users to human representatives without requiring repetition. The efficiency enables midnight questions to receive immediate answers, improving player experience regardless of when they choose to play. Beyond Basic Gambling AI powers front-end innovations beyond back-end analytics. VR-enabled casino games create immersive environments with virtual dealers and player avatars that feel surprisingly realistic. Some slot games feature AI-generated storylines or special quests, adding excitement to each spin. The technology merges with blockchain to deliver fresh experiences while maintaining verifiable fairness. AI Drives Business Metrics Of course, the implementation of AI in crypto casinos does benefit the operators. Here’s how. does Conversion Rates AI learns player habits and jumps in at just the right time. Like offering a blackjack bonus on Friday night if you're a weekend player. Or promoting a less popular slot when it sees you're into that style of game. It’s personalized, but not creepy. Revenue Growth Casinos that treat players like individuals (rather than just another wallet), tend to keep them coming back. When the platform gets your preferences, you’re more likely to stick around (and yes, deposit more). gets The Security Watchdog AI tools quietly monitor betting patterns, deposit sizes, and gameplay. It can flag the account for review if something looks off – like a sudden withdrawal spike from a normally cautious punter. What Comes Next? AI may be the missing puzzle piece for blockchain gambling, which makes online casinos safer, more engaging, and better made for each player rather than satisfying the majority. The technology promotes a healthy balance between entertainment and responsible gambling through monitoring, personalization, and new concepts. may AI will most probably continue to refine betting experiences through more precise risk assessments and immersive virtual environments. The casino industry just loves innovation, and AI proves to be an effective partner in that ongoing development process.

The Integration of AI in Blockchain-Based Gambling Platforms

Интеграцията на AI в блокчейн базирани хазартни платформи

La integració de la IA en plataformes de joc basades en blockchain

La integración de la IA en plataformas de juego basadas en blockchain

L'integrazione dell'IA nelle piattaforme di gioco basate su blockchain

ブロックチェーンベースのギャンブルプラットフォームにおけるAIの統合

Интеграцијата на вештачката интелигенција во блокчејн-базирани платформи за коцкање

Блокчейн суурилсан тоглоомын платформд AI-ийн интеграци

A integração da IA em plataformas de jogos baseadas em blockchain

Интеграция ИИ в блокчейн-платформы азартных игр

Integrimi i AI në platformat e lojërave të fatit të bazuara në Blockchain

Blockchain tabanlı kumar platformlarında AI entegrasyonu

A Kubernetes Special: Even More OpenTelemetry to Chew On

The Internet Has a Pulse, and This Sequencer Lets You Hear It

Picture this: It’s 7:30 AM on a chaotic Tuesday. While you’re scrambling to get the kids ready for school, your AI assistant quietly reorders your coffee supply, snags a discounted Air Fryer from a lightning deal, and books a table at that impossible to get into sushi spot for date night; all without you lifting a finger. This isn’t sci-fi. Meet Amazon Nova Act[1], the AI agent that actually acts. Revealed this week by Amazon’s secretive AGI Labs, Nova Act doesn’t just talk - it does. And it might just make today’s chatbots look like dial-up internet in a 5G world. From "Type This" to "Do That" Most AI tools stop at generating text or answering questions. Nova Act goes nuclear: Automates weekly Sweetgreen orders (no more sad desk salads)
Files expense reports while you sleep
Hunts down price drops like a bargain-obsessed bloodhound
Even handles password-protected tasks (RIP, "Forgot Password?" clicks) The magic? A new developer toolkit that transforms simple commands like “book,” “track,” or “compare” into multi-step actions. Want to plan a work offsite? Nova Act could theoretically: Scout Airbnb locations within budget
Compare flight prices for the team
Block calendars across time zones
Email the final itinerary All while you focus on actual work. Why This Isn’t Just Another "AI Assistant" Sure, we’ve seen attempts at action-oriented AI (cough ChatGPT Plugins cough). But Amazon claims Nova Act solves the #1 problem with automation tools: reliability. During demos, Nova Act nailed tasks competitors flubbed: ✅ 94% success rate interacting with finicky calendar widgets ✅ 91% accuracy parsing dense restaurant menus ✅ Zero accidental purchases of extended warranties (take notes, humans) “If you’re still proofreading your AI’s work, you’ve got a fancy typewriter—not an assistant,” joked Vishal Vora, an Amazon engineer, as Nova Act flawlessly scheduled his team’s PTO requests during a live demo[2]. Real World, Real Impact For individuals: Auto-negotiate Comcast bills (goodbye, hold music)


Reschedule missed doctor appointments


Batch-cancel unused subscriptions For businesses: Auto-fix common IT tickets (“Have you tried turning it off and on again?”) Monitor supply chain inventory in real-time
Scrape competitor pricing without spreadsheets Even bigger? This is Amazon’s first public step toward artificial general intelligence (AGI)—AI that can learn any task, not just preset ones. Their secret sauce? Training models across 100+ domains (travel, finance, healthcare) instead of specializing. The Catch(es) Before you quit your job to let AI run your life: Complex tasks still need training wheels - Planning a wedding? Nova Act might book the venue but forget the vegan meal options. Human oversight isn’t dead yet.
Ethical landmines - Should AI negotiate salaries? Handle medical data? Amazon’s SDK includes guardrails, but the debate’s just starting.
The “Creep Factor” - Letting AI access your bank account and email requires Olympic-level trust. Amazon assures end-to-end encryption, but skeptics will need convincing. The AI Agent Arms Race Amazon isn’t alone: OpenAI’s “Operator”: Can draft emails but struggles with actual sending
Anthropic’s Agent: Great at research, shaky at transactions
China’s $199/month “Manus”: Allegedly buys real estate (we’re skeptical) What makes Nova Act different? Scale. With Alexa in 500M+ devices, Amazon could deploy this everywhere from Echo speakers to Ring cameras. The Bottom Line Nova Act isn’t about replacing humans—it’s about eliminating the administrative sludge that consumes 60% of our workdays. The real question isn’t “Will this work?” but “What will we do with our time when it does?” As Rohit Prasad (Amazon’s AGI lead) put it: “We’re not building a better chatbot. We’re building a better day.” “What would you automate first?” References - [1] https://labs.amazon.science/blog/nova-act [2] https://www.youtube.com/watch?v=65giTqW0vC8&t=1s&themeRefresh=1 Picture this: It’s 7:30 AM on a chaotic Tuesday. While you’re scrambling to get the kids ready for school, your AI assistant quietly reorders your coffee supply, snags a discounted Air Fryer from a lightning deal, and books a table at that impossible to get into sushi spot for date night; all without you lifting a finger. This isn’t sci-fi. Meet Amazon Nova Act[1], the AI agent that actually acts. Revealed this week by Amazon’s secretive AGI Labs, Nova Act doesn’t just talk - it does. And it might just make today’s chatbots look like dial-up internet in a 5G world. From "Type This" to "Do That" Most AI tools stop at generating text or answering questions. Nova Act goes nuclear: Automates weekly Sweetgreen orders (no more sad desk salads) Files expense reports while you sleep Hunts down price drops like a bargain-obsessed bloodhound Even handles password-protected tasks (RIP, "Forgot Password?" clicks) Automates weekly Sweetgreen orders (no more sad desk salads) Files expense reports while you sleep Hunts down price drops like a bargain-obsessed bloodhound Even handles password-protected tasks (RIP, "Forgot Password?" clicks) The magic? A new developer toolkit that transforms simple commands like “book,” “track,” or “compare” into multi-step actions. Want to plan a work offsite? Nova Act could theoretically: Scout Airbnb locations within budget Compare flight prices for the team Block calendars across time zones Email the final itinerary Scout Airbnb locations within budget Compare flight prices for the team Block calendars across time zones Email the final itinerary All while you focus on actual work. Why This Isn’t Just Another "AI Assistant" Sure, we’ve seen attempts at action-oriented AI (cough ChatGPT Plugins cough). But Amazon claims Nova Act solves the #1 problem with automation tools: reliability. During demos, Nova Act nailed tasks competitors flubbed: ✅ 94% success rate interacting with finicky calendar widgets ✅ 91% accuracy parsing dense restaurant menus ✅ Zero accidental purchases of extended warranties (take notes, humans) “If you’re still proofreading your AI’s work, you’ve got a fancy typewriter—not an assistant,” joked Vishal Vora, an Amazon engineer, as Nova Act flawlessly scheduled his team’s PTO requests during a live demo[2]. Real World, Real Impact For individuals: For individuals: Auto-negotiate Comcast bills (goodbye, hold music) Reschedule missed doctor appointments Batch-cancel unused subscriptions Auto-negotiate Comcast bills (goodbye, hold music) Auto-negotiate Comcast bills (goodbye, hold music) Reschedule missed doctor appointments Reschedule missed doctor appointments Batch-cancel unused subscriptions Batch-cancel unused subscriptions For businesses: For businesses: Auto-fix common IT tickets (“Have you tried turning it off and on again?”) Monitor supply chain inventory in real-time Scrape competitor pricing without spreadsheets Monitor supply chain inventory in real-time Scrape competitor pricing without spreadsheets Even bigger? This is Amazon’s first public step toward artificial general intelligence (AGI)—AI that can learn any task, not just preset ones. Their secret sauce? Training models across 100+ domains (travel, finance, healthcare) instead of specializing. The Catch(es) Before you quit your job to let AI run your life: Complex tasks still need training wheels - Planning a wedding? Nova Act might book the venue but forget the vegan meal options. Human oversight isn’t dead yet. Ethical landmines - Should AI negotiate salaries? Handle medical data? Amazon’s SDK includes guardrails, but the debate’s just starting. The “Creep Factor” - Letting AI access your bank account and email requires Olympic-level trust. Amazon assures end-to-end encryption, but skeptics will need convincing. Complex tasks still need training wheels - Planning a wedding? Nova Act might book the venue but forget the vegan meal options. Human oversight isn’t dead yet. Complex tasks still need training wheels Ethical landmines - Should AI negotiate salaries? Handle medical data? Amazon’s SDK includes guardrails, but the debate’s just starting. Ethical landmines The “Creep Factor” - Letting AI access your bank account and email requires Olympic-level trust. Amazon assures end-to-end encryption, but skeptics will need convincing. The “Creep Factor” The AI Agent Arms Race Amazon isn’t alone: OpenAI’s “Operator”: Can draft emails but struggles with actual sending Anthropic’s Agent: Great at research, shaky at transactions China’s $199/month “Manus”: Allegedly buys real estate (we’re skeptical) OpenAI’s “Operator”: Can draft emails but struggles with actual sending OpenAI’s “Operator”: Anthropic’s Agent: Great at research, shaky at transactions Anthropic’s Agent: China’s $199/month “Manus”: Allegedly buys real estate (we’re skeptical) China’s $199/month “Manus”: What makes Nova Act different? Scale. With Alexa in 500M+ devices, Amazon could deploy this everywhere from Echo speakers to Ring cameras. The Bottom Line Nova Act isn’t about replacing humans—it’s about eliminating the administrative sludge that consumes 60% of our workdays. The real question isn’t “Will this work?” but “What will we do with our time when it does?” As Rohit Prasad (Amazon’s AGI lead) put it: “We’re not building a better chatbot. We’re building a better day.” “What would you automate first?” References - [1] https://labs.amazon.science/blog/nova-act https://labs.amazon.science/blog/nova-act https://labs.amazon.science/blog/nova-act [2] https://www.youtube.com/watch?v=65giTqW0vC8&t=1s&themeRefresh=1 https://www.youtube.com/watch?v=65giTqW0vC8&t=1s&themeRefresh=1 https://www.youtube.com/watch?v=65giTqW0vC8&t=1s&themeRefresh=1

Forget Chatbots, Meet Actionbots: Why Amazon's Nova Act Could Reshape Web Interaction

Xatırladaq ki, “Amazon”ın “Nova Act”ı web interaksiyasını dəyişdirə bilər.

Olvide los chatbots, conozca a los bots de acción: por qué la Ley Nova de Amazon podría remodelar la interacción web

चैटबॉट्स भूल जाओ, एक्शनबॉट्स से मिलें: क्यों अमेज़ॅन का नोवा अधिनियम वेब बातचीत को फिर से आकार दे सकता है

Ouvri Chatbots, Rencontre Actionbots: Poukisa Lwa Nova nan Amazon ta ka Reshape Interaction Web

Գիտեք Chatbots, Meet Actionbots: Ինչու Amazon- ի Nova Act- ը կարող է վերլուծել Web Interaction- ը

Dimenticate i chatbots, incontrate gli actionbots: perché la legge Nova di Amazon potrebbe ricostruire l'interazione web

チャットボットを忘れ、アクションボットに出会う:なぜアマゾンのノバ法がウェブインタラクションを再構築するのか

Осылайша Франклин өз беделін қайтадан қайтарып, саяси аренаға қайта

Harrojeni Chatbots, takoni Actionbots: Pse Akti Nova i Amazon-it Mund të Rishikojë Ndërveprimin në Web

I-Forget Chatbots, Meet Actionbots: Bakit ang Amazon's Nova Act ay maaaring Reshape Web Interaction

Ukukhangisa i-Chatbots, Ukukhangisa i-Actionbots: Kutheni I-Nova Act ye-Amazon ingathuthukisa ukuxhumana kwebhizinisi

Table of Links Abstract and 1 Introduction 2 LongFact: Using LLMs to generate a multi-topic benchmark for long-form factuality 3 Safe:LLM agents as factuality autoraters 4 LLMs agents can be better factuality annotators than humans 5 F1@k: Extending F1 with recall from human-preferred length 6 Larger LLMs are more factual 7 Related Work 8 Limitations 9 Conclusion, Acknowledgments, Author Contribution, and References Appendix A. Frequently asked questions B. LongFact details C. SAFE details D. Metric details E. Further analysis 7 RELATED WORK Benchmarking factuality. Recent studies on factuality in large language models have presented several factuality benchmarks. Many benchmarks such as FELM (Chen et al., 2023), FreshQA (Vu et al., 2023), and HaluEval (Li et al., 2023) simply test a model’s knowledge of various factoids. On the other hand, adversarial benchmarks such as TruthfulQA (Lin et al., 2022) and HalluQA (Cheng et al., 2023b) consist of prompts that are designed to test whether a language model will generate a false factoid learned from imitating popular misconceptions. In the long-form setting, Min et al. (2023) proposed a set of prompts that require language models to generate biographies about people, evaluated against each person’s respective Wikipedia article. Our proposed prompt set does not dispute the importance of these tasks, but rather aims to provide broad coverage of factuality in a long-form setting, whereas existing benchmarks either cover singular short factoids or only cover a small set of long-form topics. Evaluating factuality in model responses. In order to fully benchmark a model’s factuality, there must exist a reliable method for quantifying the factuality of a model response to a given prompt. Performance on factuality datasets with short ground-truth answers can often be evaluated using language models that have been fine-tuned on human evaluations (Lin et al., 2022; Sellam et al., 2020; Lin, 2004). Long-form responses, on the other hand, are harder to evaluate because they contain a much-broader range of possible factual claims. For example, FActScore (Min et al., 2023) leverages Wikipedia pages as knowledge sources for humans and language models to use when rating long-form responses on biographies about people. In reference-free contexts, RAGAS (Es et al., 2023) evaluate whether a response from a retrieval-augmented generation (RAG) system is self-contained by prompting a large language model to check the consistency among questions, retrieved contexts, answers, and statements within the answer. This matches the case study in Guan et al. (2023), which also showed that language models may serve as a strong tool for fact verification. Tian et al. (2023) leverages this intuition to demonstrate that factuality preference rankings obtained via model confidence can be used for improving factuality in responses via direct preference optimization. Improvements in human evaluation can also be important for evaluating factuality; in this domain, Cheng et al. (2023a) created an interface for human evaluation by asking fact-verification questions and checking self-consistency of model responses. Most related to our work, Chern et al. (2023) proposed a method of rating model responses by breaking them down and rating individual components via external tools and applied this in tasks such as code generation, math reasoning, and short-answer QA. Our proposed method takes inspiration from this work, but applies it in the long-form factuality setting, similar to other work that uses search-augmented language models for fact verification (Gao et al., 2023; Wang et al., 2023; Zhang & Gao, 2023). Compared to these prior methods of evaluating factuality, SAFE is automatic, does not require any model finetuning or any preset reference answers/knowledge sources, and is more-reliable than crowdsourced human annotations while only requiring a fraction of the cost. Quantifying long-form factuality. Long-form factuality is difficult to quantify because the quality of the response is influenced both by the factuality of the response (precision) and by the coverage of the response (recall). FActScore (Min et al., 2023) measures precision with respect to Wikipedia pages as knowledge sources and mentions the difficulty of measuring recall as future work. Tian et al. (2023) and Dhuliawala et al. (2023) also use precision as a proxy for a model response’s factuality. Other metrics such as token-level F1 score (Rajpurkar et al., 2016), sentence or passage-level AUC (Manakul et al., 2023), and similarity to ground-truth (Wieting et al., 2022) can account for recall as well as precision but require human responses or judgements that serve as a ground-truth. Our proposed F1@K, on the other hand, measures both precision and recall of a long-form response without requiring any preset ground-truth answers, only the number of facts from each label category. This paper is available on arxiv under CC by 4.0 Deed license. Authors:
(1) Jerry Wei, Google DeepMind and a Lead contributors;
(2) Chengrun Yang, Google DeepMind and a Lead contributors;
(3) Xinying Song, Google DeepMind and a Lead contributors;
(4) Yifeng Lu, Google DeepMind and a Lead contributors;
(5) Nathan Hu, Google DeepMind and Stanford University;
(6) Jie Huang, Google DeepMind and University of Illinois at Urbana-Champaign;
(7) Dustin Tran, Google DeepMind;
(8) Daiyi Peng, Google DeepMind;
(9) Ruibo Liu, Google DeepMind;
(10) Da Huang, Google DeepMind;
(11) Cosmo Du, Google DeepMind;
(12) Quoc V. Le, Google DeepMind. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 LongFact: Using LLMs to generate a multi-topic benchmark for long-form factuality 2 LongFact: Using LLMs to generate a multi-topic benchmark for long-form factuality 3 Safe:LLM agents as factuality autoraters 3 Safe:LLM agents as factuality autoraters 4 LLMs agents can be better factuality annotators than humans 4 LLMs agents can be better factuality annotators than humans 5 F1@k: Extending F1 with recall from human-preferred length 5 F1@k: Extending F1 with recall from human-preferred length 6 Larger LLMs are more factual 6 Larger LLMs are more factual 7 Related Work 7 Related Work 8 Limitations 8 Limitations 9 Conclusion, Acknowledgments, Author Contribution, and References 9 Conclusion, Acknowledgments, Author Contribution, and References Appendix Appendix A. Frequently asked questions A. Frequently asked questions B. LongFact details B. LongFact details C. SAFE details C. SAFE details D. Metric details D. Metric details E. Further analysis E. Further analysis 7 RELATED WORK Benchmarking factuality. Recent studies on factuality in large language models have presented several factuality benchmarks. Many benchmarks such as FELM (Chen et al., 2023), FreshQA (Vu et al., 2023), and HaluEval (Li et al., 2023) simply test a model’s knowledge of various factoids. On the other hand, adversarial benchmarks such as TruthfulQA (Lin et al., 2022) and HalluQA (Cheng et al., 2023b) consist of prompts that are designed to test whether a language model will generate a false factoid learned from imitating popular misconceptions. In the long-form setting, Min et al. (2023) proposed a set of prompts that require language models to generate biographies about people, evaluated against each person’s respective Wikipedia article. Our proposed prompt set does not dispute the importance of these tasks, but rather aims to provide broad coverage of factuality in a long-form setting, whereas existing benchmarks either cover singular short factoids or only cover a small set of long-form topics. Benchmarking factuality. Evaluating factuality in model responses. In order to fully benchmark a model’s factuality, there must exist a reliable method for quantifying the factuality of a model response to a given prompt. Performance on factuality datasets with short ground-truth answers can often be evaluated using language models that have been fine-tuned on human evaluations (Lin et al., 2022; Sellam et al., 2020; Lin, 2004). Long-form responses, on the other hand, are harder to evaluate because they contain a much-broader range of possible factual claims. For example, FActScore (Min et al., 2023) leverages Wikipedia pages as knowledge sources for humans and language models to use when rating long-form responses on biographies about people. In reference-free contexts, RAGAS (Es et al., 2023) evaluate whether a response from a retrieval-augmented generation (RAG) system is self-contained by prompting a large language model to check the consistency among questions, retrieved contexts, answers, and statements within the answer. This matches the case study in Guan et al. (2023), which also showed that language models may serve as a strong tool for fact verification. Tian et al. (2023) leverages this intuition to demonstrate that factuality preference rankings obtained via model confidence can be used for improving factuality in responses via direct preference optimization. Improvements in human evaluation can also be important for evaluating factuality; in this domain, Cheng et al. (2023a) created an interface for human evaluation by asking fact-verification questions and checking self-consistency of model responses. Most related to our work, Chern et al. (2023) proposed a method of rating model responses by breaking them down and rating individual components via external tools and applied this in tasks such as code generation, math reasoning, and short-answer QA. Our proposed method takes inspiration from this work, but applies it in the long-form factuality setting, similar to other work that uses search-augmented language models for fact verification (Gao et al., 2023; Wang et al., 2023; Zhang & Gao, 2023). Compared to these prior methods of evaluating factuality, SAFE is automatic, does not require any model finetuning or any preset reference answers/knowledge sources, and is more-reliable than crowdsourced human annotations while only requiring a fraction of the cost. Evaluating factuality in model responses. Quantifying long-form factuality. Long-form factuality is difficult to quantify because the quality of the response is influenced both by the factuality of the response (precision) and by the coverage of the response (recall). FActScore (Min et al., 2023) measures precision with respect to Wikipedia pages as knowledge sources and mentions the difficulty of measuring recall as future work. Tian et al. (2023) and Dhuliawala et al. (2023) also use precision as a proxy for a model response’s factuality. Other metrics such as token-level F1 score (Rajpurkar et al., 2016), sentence or passage-level AUC (Manakul et al., 2023), and similarity to ground-truth (Wieting et al., 2022) can account for recall as well as precision but require human responses or judgements that serve as a ground-truth. Our proposed F1@K, on the other hand, measures both precision and recall of a long-form response without requiring any preset ground-truth answers, only the number of facts from each label category. Quantifying long-form factuality. 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 available on arxiv Authors: (1) Jerry Wei, Google DeepMind and a Lead contributors; (2) Chengrun Yang, Google DeepMind and a Lead contributors; (3) Xinying Song, Google DeepMind and a Lead contributors; (4) Yifeng Lu, Google DeepMind and a Lead contributors; (5) Nathan Hu, Google DeepMind and Stanford University; (6) Jie Huang, Google DeepMind and University of Illinois at Urbana-Champaign; (7) Dustin Tran, Google DeepMind; (8) Daiyi Peng, Google DeepMind; (9) Ruibo Liu, Google DeepMind; (10) Da Huang, Google DeepMind; (11) Cosmo Du, Google DeepMind; (12) Quoc V. Le, Google DeepMind. Authors: Authors: (1) Jerry Wei, Google DeepMind and a Lead contributors; (2) Chengrun Yang, Google DeepMind and a Lead contributors; (3) Xinying Song, Google DeepMind and a Lead contributors; (4) Yifeng Lu, Google DeepMind and a Lead contributors; (5) Nathan Hu, Google DeepMind and Stanford University; (6) Jie Huang, Google DeepMind and University of Illinois at Urbana-Champaign; (7) Dustin Tran, Google DeepMind; (8) Daiyi Peng, Google DeepMind; (9) Ruibo Liu, Google DeepMind; (10) Da Huang, Google DeepMind; (11) Cosmo Du, Google DeepMind; (12) Quoc V. Le, Google DeepMind.

A Smarter Way to Check If AI Answers Are Correct

Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 LongFact: Using LLMs to generate a multi-topic benchmark for long-form factuality 2 LongFact: Using LLMs to generate a multi-topic benchmark for long-form factuality 3 Safe:LLM agents as factuality autoraters 3 Safe:LLM agents as factuality autoraters 4 LLMs agents can be better factuality annotators than humans 4 LLMs agents can be better factuality annotators than humans 5 F1@k: Extending F1 with recall from human-preferred length 5 F1@k: Extending F1 with recall from human-preferred length 6 Larger LLMs are more factual 6 Larger LLMs are more factual 7 Related Work 7 Related Work 8 Limitations 8 Limitations 9 Conclusion, Acknowledgments, Author Contribution, and References 9 Conclusion, Acknowledgments, Author Contribution, and References Appendix Appendix A. Frequently asked questions A. Frequently asked questions B. LongFact details B. LongFact details C. SAFE details C. SAFE details D. Metric details D. Metric details E. Further analysis E. Further analysis 3 SAFE: LLM AGENTS AS FACTUALITY AUTORATERS A key obstacle in benchmarking long-form factuality is that a benchmark requires a reliable method of evaluating model responses. Existing automated evaluation methods such as BLEURT (Sellam et al., 2020), ROUGE (Lin, 2004; Ganesan, 2018), and prompting language models (Min et al., 2023; Vu et al., 2023) operate by comparing a model response to a preset reference answer or knowledge source. This works well for prompts that only require a short-answer response, as there is a definite and singular correct answer. However, this setting is suboptimal for prompts that require a long-form response, as it is difficult to compile predetermined answers/knowledge that comprehensively and nonredundantly cover all facts that a model response may contain. Instead, we propose that there are two key insights that are crucial to accurately evaluate factuality in a long-form setting. First, we agree with the idea that long-form responses should be evaluated at the granularity of individual facts, following Min et al. (2023). This is more fine-grained than the granularity of sentences, as a single sentence may contain multiple individual facts (see Figure 1 for an example), and it is essential for optimal performance since isolating each individual fact allows for precise and focused evaluation. Second, in lieu of predetermined reference answers or knowledge sources, Google Search can be leveraged to obtain dynamic reference answers for each fact in a response with the best tradeoff among accuracy, accessibility, and timeliness (Augenstein et al., 2023). This ensures that evaluation is done with respect to reference answers that specifically cover the particular set of facts that a long-form response can contain. For these reasons, we propose Search-Augmented Factuality Evaluator (SAFE),[4] which joins these insights by using a language model to (a) split a long-form response into individual self-contained facts, (b) determine whether each individual fact is relevant to answering the prompt in the context of the response, and (c) for each relevant fact, iteratively issue Google Search queries in a multi-step process and reason about whether the search results support or do not support the fact. We consider the key innovation in SAFE to be using a language model as an agent to generate multi-step Google Search queries and carefully reason about whether facts are supported by search results (Figure 3 and Appendix C.2 shows examples of these reasoning chains). As shown in Figure 1, to split a long-form response into individual self-contained facts, we first prompt the language model to split each sentence in a long-form response into individual facts. We then revise each individual fact to be self-contained by instructing the model to replace vague references (e.g., pronouns) with the proper entities that they are referring to in the context of the response. To rate each self-contained individual fact, we use the language model to reason about whether the fact is relevant to answering the prompt in the context of the response.[5] Each remaining relevant fact is then rated as “supported” or “not supported” using a multi-step approach. In each step, the model generates a search query based on the fact to rate and the search results that have previously been obtained. After a set number of steps, the model performs reasoning to determine whether the fact is supported by the search results, as shown in Figure 3. Once all facts are rated, the outputted metrics from SAFE for a given prompt–response pair are the number of “supported” facts, the number of “irrelevant” facts, and the number of “not-supported” facts. Our detailed implementation of SAFE is shown in Appendix C.1, and SAFE is publicly available at https://github.com/ google-deepmind/long-form-factuality/tree/main/eval/safe. 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 available on arxiv [4] SAFE is similar to and inspired by prior work such as Gao et al. (2023) and Wang et al. (2023). [5] This identifies statements that should not be rated (e.g., “I don’t know the answer to that”). Authors:
(1) Jerry Wei, Google DeepMind and a Lead contributors;
(2) Chengrun Yang, Google DeepMind and a Lead contributors;
(3) Xinying Song, Google DeepMind and a Lead contributors;
(4) Yifeng Lu, Google DeepMind and a Lead contributors;
(5) Nathan Hu, Google DeepMind and Stanford University;
(6) Jie Huang, Google DeepMind and University of Illinois at Urbana-Champaign;
(7) Dustin Tran, Google DeepMind;
(8) Daiyi Peng, Google DeepMind;
(9) Ruibo Liu, Google DeepMind;
(10) Da Huang, Google DeepMind;
(11) Cosmo Du, Google DeepMind;
(12) Quoc V. Le, Google DeepMind. Authors: Authors: (1) Jerry Wei, Google DeepMind and a Lead contributors; (2) Chengrun Yang, Google DeepMind and a Lead contributors; (3) Xinying Song, Google DeepMind and a Lead contributors; (4) Yifeng Lu, Google DeepMind and a Lead contributors; (5) Nathan Hu, Google DeepMind and Stanford University; (6) Jie Huang, Google DeepMind and University of Illinois at Urbana-Champaign; (7) Dustin Tran, Google DeepMind; (8) Daiyi Peng, Google DeepMind; (9) Ruibo Liu, Google DeepMind; (10) Da Huang, Google DeepMind; (11) Cosmo Du, Google DeepMind; (12) Quoc V. Le, Google DeepMind.

SAFE: A New AI Tool for Fact-Checking Long-Form Responses

Authors:
(1) Jinyu Cai, Waseda University (bluelink@toki.waseda.jp);
(2) Jialong Li, Waseda University (lijialong@fuji.waseda.jp);
(3) Mingyue Zhang, Southwest University (myzhangswu@swu.edu.cn);
(4) Munan Li, Dalian Maritime University (limunan@dlmu.edu.cn);
(5) Chen-Shu Wang, National Taipei University of Technology (wangcs@ntut.edu.tw);
(6) Kenji Tei, Tokyo Institute of Technology (tei@c.titech.ac.jp). Authors: Authors: (1) Jinyu Cai, Waseda University (bluelink@toki.waseda.jp); (2) Jialong Li, Waseda University (lijialong@fuji.waseda.jp); (3) Mingyue Zhang, Southwest University (myzhangswu@swu.edu.cn); (4) Munan Li, Dalian Maritime University (limunan@dlmu.edu.cn); (5) Chen-Shu Wang, National Taipei University of Technology (wangcs@ntut.edu.tw); (6) Kenji Tei, Tokyo Institute of Technology (tei@c.titech.ac.jp). Table of Links Abstract and I. Introduction Abstract and I. Introduction II. Background and Related Work II. Background and Related Work III. Framework Design III. Framework Design IV. Evaluation IV. Evaluation V. Conclusion and Future Work, Acknowledgement, and References V. Conclusion and Future Work, Acknowledgement, and References III. FRAMEWORK DESIGN A. Overview A. Overview In this section, we offer a detailed overview of our system, as depicted in Figure 1. This figure provides a visual representation of our framework, highlighting its key components and their interrelationships. Our system is primarily composed of two types of agents: the Supervisor, tasked with enforcing established guidelines, and the Participant, whose goal is to convey specific, human-defined information discreetly. Participants must dynamically refine their communication approaches, drawing from past dialogues, to transmit information effectively while remaining undetected. In the entire system, the actions of both participants and the supervisor are driven by the LLM. Initially, we establish the foundational information for each agent, including role setting, background knowledge, and primary tasks. Subsequently, the participant agents engage in dialogues with each other. After each dialogue turn, the supervisory agent reviews the conversation to determine if any pre-set rules have been violated. In cases of rule violation, the supervisor interrupts the dialogue, providing feedback about the infringing text and the rationale behind it. Throughout this process, the dialogues between participants, along with the supervisory feedback on violations, are recorded separately in the “Dialogue History” and “Violation Log.” Before new dialogues, participant agents use the Reflection Module to develop or refine “Regulations” from the Violation Log, guiding their dialogue creation. Successful dialogues without detection proceed to an interview phase for perspective assessment. The Reflection Module then reevaluates these insights, generating or enhancing “Guidance” for future dialogues. The Planning Module activates for more direct dialogue content guidance whenever Regulations or Guidance are updated. B. Participant Agents B. Participant Agents Participant agents in our system are composed of several modules, including Memory, Dialogue, Reflection, and Summary, all powered by LLMs. To increase the system’s flexibility and minimize redundancy, we’ve structured the prompts for each module around seven primary elements: “Background Information,” “Dialogue History,” “Violation Log,” “Regulations,” “Guidance,” “Plan,” and “Instructions.” “Background Information” delivers essential data and objectives pertinent to the experimental setup. The Memory module manages “Dialogue History” and “Violation Log,” which respectively track participant dialogues and instances of detection by the supervisor. Overcoming the challenge of effectively communicating regulated topics under supervision tests the linguistic prowess of LLMs. To address this, we’ve integrated “Regulations,” “Guidance,” and “Plan” as crucial components, formulated by the Reflection and Summary modules, to assist agents in stealthily disseminating information. “Instructions” set specific tasks for the LLM within each module. 1) Dialogue Module: This module generates dialogue content based on short-term dialogue records. Extracting historical dialogue information from the Memory module, it inputs this into the LLM to understand and construct sentences that evade supervision while effectively transmitting information. In the Dialogue module’s prompt, “Background Information” provides necessary background, while “Plan” guide the achievement of objectives. “Instructions” offer LLM-specific execution directives. 1) Dialogue Module: 2) Memory Module: The Memory module stores all dialogue history and records detected by the supervisor, Specifically, it comprises three parts: background Information, dialogue history, and violation records. The background information includes role settings, experimental background knowledge, and global objectives. “Dialogue History” and “Violation Log” save dialogue records and past detections by the Supervisor, respectively. “Background information” and “Dialogue History” serves as short-term [36] memory, containing only the current round’s dialogue. “Violation Log,” as long-term memory, records violations from each evolutionary round. When dialogues are detected by the Supervisor, relevant feedback is added to the “Violation Log,” triggering a new evolutionary process. 2) Memory Module: 2) Memory Module: Excessive memory information can potentially distract the LLM and lead to a decline in performance. Hence, the “Dialogue History” and “Violation Log” in the memory module are regularly maintained. This involves inputting earlier memories and employing the LLM to distill crucial information from these records, effectively compressing and consolidating them. This approach not only preserves essential historical data but also optimizes the LLM’s performance, striking a balance between comprehensive memory retention and efficient processing. 3) Reflection Module: The Reflection Module is activated at the beginning and end of each dialogue evolution cycle, with its core purpose being the generation of improved strategies based on historical records. At the start of an evolution cycle, the module utilizes the “Violation Log” as its input to analyze past failures and, based on these insights, formulates “Regulations” aimed at effectively circumventing supervision in future dialogues. At the cycle’s end, the Reflection Module is reactivated, turning its focus to the “Dialogue History”. This step is crucial for reviewing and assessing the dialogue content to verify the successful completion of the primary task of information conveyance. If the module identifies deficiencies in information delivery or objectives not met, it then proposes “Guidance” for subsequent dialogues, thereby enhancing the agent’s capacity for information transmission. This design establishes the Reflection Module as a key self-evaluation and strategy adjustment mechanism within the system, ensuring continuous improvement and adaptability of the dialogue system in a dynamically changing regulatory environment. 3) Reflection Module: 3) Reflection Module: To fulfill these requirements, the LLM must possess a high degree of reasoning ability, capable of inferring the supervisor’s criteria from failure records and identifying communication deficiencies from dialogue history, thereby formulating appropriate strategies and improvement suggestions. To enhance the LLM’s reasoning capabilities, the design of the prompt incorporates the effective and cost-efficient Chain of Thought [37] (COT) method as the guiding principle for the reflection and planning modules. This approach guides the LLM in conducting criterion analysis and, in combination with the main task, generating “Regulations” and “Guidance”. Specifically, within the Reflection Module, the prompt includes “Background Information,” “Dialogue History” or “Violation Log,” “Old Guidance” or ”Old Regulations,” and “Instructions.” The “Instructions” first prompt the LLM to engage in preliminary thinking, for instance: “Please infer what kind of guidelines the Supervisor is following from the Violation Log.” This is followed by a conclusion question: “Based on this information, update existing regulations to better avoid supervision.” The content returned by the LLM will serve as the new regulation for the next round of dialogue. 4) Planning Module: The facet of “Regulations” is centered on devising strategies that sidestep supervision, whereas “Guidance” is all about the effective relay of information. Although both are pivotal in managing dialogues, they might clash in certain supervised scenarios, thereby impacting the language’s overall proficiency and efficiency [38]. In light of this, integrating the Planning Module becomes indispensable. 4) Planning Module: 4) Planning Module: Within the Planning Module, the prompt includes “Background Information,” “Regulations,” “Guidance,” and “Instructions”. In the “Instructions,” we mandate the LLM to use “Regulations” as a baseline. Building upon this foundation, it is guided by the “Guidance” to formulate a tangible and actionable “Plan.” This approach ensures the delivery of clear and effective strategies for future dialogues. C. Supervisory Agent C. Supervisory Agent The role of the Supervisor is designed as an abstract concept, primarily to simulate social network supervision. This design reflects an important observation: in social networks, content supervision often focuses on direct judgment of specific information, rarely influenced by extensive or lengthy contextual information. Thus, unlike participants with more complex background information, the Supervisor’s function and role are simplified here to highlight its behavior characteristics in actual social network supervision. The Supervisor’s task mainly focuses on assessing whether content complies with certain standards or rules, a relatively straightforward and well-defined process. To mimic the existing review mechanisms of platforms, which typically combine keyword filters with “human” oversight, the Supervisor initially employs keyword filtering for a preliminary review of the dialogue content. Content that passes this initial screening is then subjected to further evaluation by the LLM. The prompt for the Supervisor includes just two components: “Dialogue History” and “Instructions.” “Dialogue History“ comprises the content of the dialogue exchanged between participant agents in that particular round, and “Instructions“ outline the criteria and guidelines that the v D. Similarities and Differences between Our Framework and Evolutionary Computing D. Similarities and Differences between Our Framework and Evolutionary Computing It should be noted that the simulation framework proposed in this paper is similar to evolutionary computing in some aspects, but there are also significant differences. The similarities include: (i) In evolutionary computing, individuals need to adapt to environmental pressures for survival and reproduction. Similarly, participants in this framework need to adapt to supervisory pressures and adjust their strategies for effective information transmission; (ii) The Reflection and Summary modules generate a “new generation” by analyzing past dialogues and violation records (i.e., records of lowfitness individuals), similar to the repeated iteration process in evolutionary computing; (iii) Since the generation of LLMs inherently involves randomness, the process of using LLMs to generate the next generation includes a de facto introduction of random mutations; (iv) In the Reflection and Memory modules, we prioritize past records, akin to the “selection” process, where individuals with higher fitness have greater weight in the generation of the new generation. The main differences stem from the particularities of “language expression”, making it infeasible to directly apply traditional evolutionary computing algorithms (such as genetic algorithms and genetic programming). They are: (i) The generation strategy of language text is difficult to encode and to perform operations of natural selection, genetic mutation, and crossover; (ii) Evolutionary computing often aims at finding the optimal solution for a specific problem environment, however, in the problem setting of this paper, it is difficult to define an explicit fitness function to evaluate what strategy is “optimal”. 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

Your Social Media Supervisor Could Soon Be an AI, And It's Getting Smarter

Hello JavaScript Enthusiasts! Welcome to a new edition of "This Week in JavaScript"! This Week in JavaScript Today, we're covering a critical Next.js security patch you can't afford to miss, OpenAI's revolutionary 4o Image Generation, and Vue.js's impressive growth stats—plus exciting new tools that'll supercharge your development workflow! Next.js Next.js Middleware Security Vulnerability Next.js Middleware Security Vulnerability Next.js just patched a critical vulnerability (CVE-2025-29927) affecting all self-hosted applications with output: 'standalone'. Next.js just patched a critical vulnerability Next.js just patched a critical vulnerability Why It Matters: Why It Matters: Authentication Bypass: Attackers could completely skip middleware-based authorization checks
All Versions Affected: Every major version prior to the patches requires updates
Simple Exploit: Using a spoofed header allows bypassing all security middleware Authentication Bypass: Attackers could completely skip middleware-based authorization checks Authentication Bypass: All Versions Affected: Every major version prior to the patches requires updates All Versions Affected: Simple Exploit: Using a spoofed header allows bypassing all security middleware Simple Exploit: If you're running a self-hosted Next.js app, update immediately to the patched versions: 15.x: fixed in 15.2.3
14.x: fixed in 14.2.25
13.x: fixed in 13.5.9
12.x: fixed in 12.3.5 15.x: fixed in 15.2.3 15.x: 14.x: fixed in 14.2.25 14.x: 13.x: fixed in 13.5.9 13.x: 12.x: fixed in 12.3.5 12.x: OpenAI's 4o Image Generation OpenAI's 4o Image Generation OpenAI has integrated its most advanced image generator directly into the GPT-4o language model. However, as of now, this feature is not publicly available through the API but is expected to be rolled out gradually in the near future. most advanced image generator most advanced image generator Key Features: Key Features: Native Integration: Built directly into the language model, not a separate tool
Text Rendering: Perfect for creating diagrams, charts, and text-heavy visuals
Multi-turn Refinement: Natural conversation to improve images incrementally
Knowledge-Backed Generation: Leverages the model's understanding for accurate visuals Native Integration: Built directly into the language model, not a separate tool Native Integration: Text Rendering: Perfect for creating diagrams, charts, and text-heavy visuals Text Rendering: Multi-turn Refinement: Natural conversation to improve images incrementally Multi-turn Refinement: Knowledge-Backed Generation: Leverages the model's understanding for accurate visuals Knowledge-Backed Generation: This isn't just about pretty pictures—it's about creating useful visuals that communicate meaning effectively through simple conversation. Use Cases in JavaScript & Web Development: Use Cases in JavaScript & Web Development: AI-Generated Assets in Web Apps: React/Vue-based UIs can dynamically generate images based on user input.
Game Development & WebGL: AI-powered textures and procedural assets for Three.js, Babylon.js, and PlayCanvas.
Generative Art & Design: AI-assisted image manipulation in p5.js or Processing.js.
AI-Powered E-commerce & Marketing: Dynamic product visuals and personalized ads.
Automated Data Visualization: AI-generated infographics and reports in D3.js or Chart.js.
Web-Based Image Editing Tools: AI-enhanced photo editing and filters using the Canvas API & TensorFlow.js.
Interactive Learning Platforms: AI-created diagrams and illustrations for educational web apps. AI-Generated Assets in Web Apps: React/Vue-based UIs can dynamically generate images based on user input. AI-Generated Assets in Web Apps: Game Development & WebGL: AI-powered textures and procedural assets for Three.js, Babylon.js, and PlayCanvas. Game Development & WebGL: Generative Art & Design: AI-assisted image manipulation in p5.js or Processing.js. Generative Art & Design: AI-Powered E-commerce & Marketing: Dynamic product visuals and personalized ads. AI-Powered E-commerce & Marketing: Automated Data Visualization: AI-generated infographics and reports in D3.js or Chart.js. Automated Data Visualization: Web-Based Image Editing Tools: AI-enhanced photo editing and filters using the Canvas API & TensorFlow.js. Web-Based Image Editing Tools: Interactive Learning Platforms: AI-created diagrams and illustrations for educational web apps. Interactive Learning Platforms: State of Vue.js Report 2025 State of Vue.js Report 2025 The latest Vue.js report shows impressive growth and ecosystem shifts that every frontend developer should know about. latest Vue.js report latest Vue.js report Striking Statistics: Striking Statistics: 93.4% Developer Loyalty: Up from 90% in 2021 planning to use Vue for their next project
80% Pinia Dominance: The state management landscape has completely transformed
38.4% Vuex Usage: The former champion has seen a dramatic decline
TypeScript Explosion: Usage surged from 38% (third place) to 82% (first place) 93.4% Developer Loyalty: Up from 90% in 2021 planning to use Vue for their next project 93.4% Developer Loyalty: 80% Pinia Dominance: The state management landscape has completely transformed 80% Pinia Dominance: 38.4% Vuex Usage: The former champion has seen a dramatic decline 38.4% Vuex Usage: TypeScript Explosion: Usage surged from 38% (third place) to 82% (first place) TypeScript Explosion: This dramatic shift shows Vue's continued momentum and the JavaScript community's growing embrace of type safety. Tools & Releases You Should Know About Tools & Releases You Should Know About Let's speed-run through some of the other big tool updates this week! Bun v1.2.7: Introduces a new CookieMap API for simplified cookie handling, improved TypeScript declarations removing Node.js/DOM conflicts, and includes 35 bug fixes improving stability and compatibility.


pnpm 10.7: Now lets you patch dependencies by version ranges for granular control, adds environment variables support in workspace config, and enhances configuration visibility across workspace files.


Babel 7.27.0: Brings default support for correct import attributes syntax, better alignment with standard JavaScript and tooling, and improved TypeScript preset behavior.


Babylon.js 8.0: Delivers Image-Based Lighting shadows for realistic rendering, Area Lights for movie-set style lighting effects, a Node Render Graph for complete pipeline customization, and a Lightweight Viewer for minimal 3D embedding.


Lexical 0.29: Meta's extensible text editor framework featuring accessible, cross-platform design (web and iOS) and a minimalist approach with plugin-based expansion. Bun v1.2.7: Introduces a new CookieMap API for simplified cookie handling, improved TypeScript declarations removing Node.js/DOM conflicts, and includes 35 bug fixes improving stability and compatibility. Bun v1.2.7: Introduces a new CookieMap API for simplified cookie handling, improved TypeScript declarations removing Node.js/DOM conflicts, and includes 35 bug fixes improving stability and compatibility. Bun v1.2.7: Bun v1.2.7: pnpm 10.7: Now lets you patch dependencies by version ranges for granular control, adds environment variables support in workspace config, and enhances configuration visibility across workspace files. pnpm 10.7: Now lets you patch dependencies by version ranges for granular control, adds environment variables support in workspace config, and enhances configuration visibility across workspace files. pnpm 10.7: pnpm 10.7: Babel 7.27.0: Brings default support for correct import attributes syntax, better alignment with standard JavaScript and tooling, and improved TypeScript preset behavior. Babel 7.27.0: Brings default support for correct import attributes syntax, better alignment with standard JavaScript and tooling, and improved TypeScript preset behavior. Babel 7.27.0: Babel 7.27.0: Babylon.js 8.0: Delivers Image-Based Lighting shadows for realistic rendering, Area Lights for movie-set style lighting effects, a Node Render Graph for complete pipeline customization, and a Lightweight Viewer for minimal 3D embedding. Babylon.js 8.0: Delivers Image-Based Lighting shadows for realistic rendering, Area Lights for movie-set style lighting effects, a Node Render Graph for complete pipeline customization, and a Lightweight Viewer for minimal 3D embedding. Babylon.js 8.0: Babylon.js 8.0: Lexical 0.29: Meta's extensible text editor framework featuring accessible, cross-platform design (web and iOS) and a minimalist approach with plugin-based expansion. Lexical 0.29: Meta's extensible text editor framework featuring accessible, cross-platform design (web and iOS) and a minimalist approach with plugin-based expansion. Lexical 0.29: Lexical 0.29: And that's it for the twenty-eighth issue of "This Week in JavaScript", brought to you by jam.dev—the tool that makes it impossible for your team to send you bad bug reports. This Week in JavaScript jam.dev jam.dev Feel free to share this newsletter with a fellow developer, and make sure you're subscribed to get notified about the next issue. Until next time, happy coding!

Next.js Locks the Back Door, OpenAI Unlocks Image Generation, and Vue Steals the Spotlight

Expertise Delivered Frontend, Backend & QA Industry Insurance The Client Our client, HealthPro Insurance, ranks among the top five Medicare insurance brokers in the United States. Certified by Healthcare.gov, they specialize in offering guidance with affordable health insurance and support for Dental, Vision, and long-term life insurance. Project Scope Our client employed agents to sell its policies like any other insurance company. The policies are sold and managed through a government-approved website. Our client had to manually refresh each agent's record, export the files to the admin account, and upload them into their CRM software daily, a highly inefficient process. This highly time-consuming process required an automated solution to reduce their manual efforts in collecting and uploading their everyday agents’ sales records to their CRM. The primary objective of this project is to introduce automation to free up valuable resources and time with policy data management. Subsequently, it would also enhance the efficiency and accuracy of customer data within our client’s CRM. The Challenge Uploading customer data to the CRM manually took 10-12 hours, leading to missed uploads and technical issues. The process also observed compromises due to technical problems and human errors. Such mishaps prevented the latest customer data from reflecting on client-facing portals, negatively affecting the company’s reputation and increasing the churn rate. Additionally, for instance, if a single web page listed 100 agents and there were 40 web pages, hitting the ‘refresh button’ for all 100 agents on a single page could cause the site to malfunction. Our client wanted to address the following challenges and roadblocks hindering their data management process. Significantly reduce the total time to migrate customer data to the CRM.
Eliminating the agent’s time invested in conducting repetitive tasks that could have been utilized in selling policies.
Mitigate errors while refreshing the agent data tab, such as' 429 rate limit’ or ‘Unable to load page. Significantly reduce the total time to migrate customer data to the CRM. Eliminating the agent’s time invested in conducting repetitive tasks that could have been utilized in selling policies. Mitigate errors while refreshing the agent data tab, such as' 429 rate limit’ or ‘Unable to load page. Why Maruti Techlabs? Our client had previously partnered with Maruti Techlabs to migrate their services from the public to the private cloud to improve their servers' and databases' stability and security. Furthermore, a consultant recommended us after observing that we had developed a similar solution for one of our other clients in the same industry. They recognized our expertise in scalable cloud infrastructure and customized software solutions and collaborated with us to optimize their insurance processes. cloud infrastructure customized software solutions Building on our successful track record and the endorsement of a trusted consultant, the client chose to extend our partnership for another project. Solution Our client's primary objective was to refresh the ‘Payment’ column for different agents from the admin login. To achieve this goal, we created a Selenium script that would log in through the Admin’s credentials and refresh each row that doesn’t reflect the status ‘As of Today.’ The script will scrape data from the website, select a date range for which it needs to export data and send it to the admin email. The records will be exported in a CSV file. Our implemented script will then download and upload the customer records to the Vtiger CRM. All the records can be migrated to our client’s CRM daily, weekly, or monthly, as they require. To design a cost-effective solution, we implemented AWS Batch as a service, where each script runs independently but simultaneously. Once it concludes its execution, the AWS batch stops, saving money. We have established an ongoing partnership with our client. In addition to the script implemented above, our experts have brainstormed ways to enhance the solution's effectiveness. Our strategic executions for future enhancements include: Script Automation - We plan to introduce automation as the script needs manual activation.
Memory Function - This would allow our script to begin execution exactly where it left off if it breaks due to issues like' 429 rate limit’ or ‘Page not loading properly.’
Preset Script Termination - This feature automatically stops executing all scripts at a predetermined time to avoid logging issues.
Data/Metrics Dashboard - Our client could observe daily, weekly, and monthly sales data and reports in real-time. Script Automation - We plan to introduce automation as the script needs manual activation. Script Automation - Memory Function - This would allow our script to begin execution exactly where it left off if it breaks due to issues like' 429 rate limit’ or ‘Page not loading properly.’ Memory Function - Preset Script Termination - This feature automatically stops executing all scripts at a predetermined time to avoid logging issues. Preset Script Termination - Data/Metrics Dashboard - Our client could observe daily, weekly, and monthly sales data and reports in real-time. Data/Metrics Dashboard - Communication and Collaboration To optimize the execution of this project and ensure timely delivery, we deployed a dedicated team comprising: 1 Backend Developer


1 Quality Assurance Engineer


1 DevOps Engineer 1 Backend Developer 1 Backend Developer 1 Quality Assurance Engineer 1 Quality Assurance Engineer 1 DevOps Engineer 1 DevOps Engineer We extensively communicated with our client's Chief of Sales throughout the project, using various platforms for timely communication. Slack - Quick communication,
Google Meet - Daily updates and queries,
Zoom - Weekly roll-outs
Gmail - Official channel to share contracts, weekly updates, and more.
Jira - Project management & issue tracking
Confluence - Assembling plans and documents behind the solution. Slack - Quick communication, Google Meet - Daily updates and queries, Zoom - Weekly roll-outs Gmail - Official channel to share contracts, weekly updates, and more. Jira - Project management & issue tracking Confluence - Assembling plans and documents behind the solution. Technology Stack Result Our solution significantly improved our client’s policy data collection process. Here are some significant benefits they observed. The real-time status of policies in the CRM is now readily available.


We maximized agent productivity in customer outreach and sales by automating clerical work.


CRM record updates are now completed 75% faster.


Automated scripts have significantly enhanced customer data accuracy and eliminated the risk of duplication.


Timely access to customer data has prevented business loss, thereby increasing customer satisfaction.


Significant cost savings have been achieved through AWS batch as a service. The real-time status of policies in the CRM is now readily available. The real-time status of policies in the CRM is now readily available. We maximized agent productivity in customer outreach and sales by automating clerical work. We maximized agent productivity in customer outreach and sales by automating clerical work. CRM record updates are now completed 75% faster. CRM record updates are now completed 75% faster. Automated scripts have significantly enhanced customer data accuracy and eliminated the risk of duplication. Automated scripts have significantly enhanced customer data accuracy and eliminated the risk of duplication. Timely access to customer data has prevented business loss, thereby increasing customer satisfaction. Timely access to customer data has prevented business loss, thereby increasing customer satisfaction. Significant cost savings have been achieved through AWS batch as a service. Significant cost savings have been achieved through AWS batch as a service. Maruti Techlabs’ collaborative efforts have successfully streamlined our client’s customer data collection process. As we continue our ongoing partnership, we remain committed to further optimizing operations and driving continued success. Our Development Process We follow Agile, Lean, & DevOps best practices to create a superior prototype that brings your users’ ideas to fruition through collaboration & rapid execution. Our top priority is quick reaction time & accessibility. We really want to be your extended team, so apart from the regular meetings, you can be sure that each of our team members is one phone call, email, or message away.

How Maruti Techlabs Simplified Medicare Policy Sales for HealthPro Insurance, Cutting Processing Time by 50%

Hoe Maruti Techlabs Medicare-polisverkope vir HealthPro-versekering vereenvoudig het, verwerkingstyd met 50% verminder het

ማሩቲ ቴክላብስ እንዴት የሜዲኬር ፖሊሲን ለ HealthPro ኢንሹራንስ እንዳቀለለ፣ የሂደት ጊዜን በ 50% መቀነስ

Як Maruti Techlabs спрасціла продаж полісаў Medicare для страхавання HealthPro, скараціўшы час апрацоўкі на 50%

Как Maruti Techlabs опрости продажбите на полици за Medicare за застраховка HealthPro, намалявайки времето за обработка с 50%

Cómo Maruti Techlabs simplificó las ventas de pólizas de Medicare para HealthPro Insurance, reduciendo el tiempo de procesamiento en un 50 %

Maruti Techlabs が HealthPro Insurance のメディケア ポリシー販売を簡素化し、処理時間を 50% 短縮した方法

როგორ გაამარტივა Maruti Techlabs-მა მკურნალი პოლიტიკის გაყიდვები HealthPro დაზღვევისთვის, დამუშავების დროის შემცირება 50%-ით

Ahoana ny fomba nanatsorana ny fivarotana politikan'ny Medicare ho an'ny fiantohana HealthPro i Maruti Techlabs, manapaka 50% ny fotoana fanodinana

Kamoo Maruti Techlabs e Nolofaditšego Thekišo ya Pholisi ya Medicare bakeng sa Inšorense ya HealthPro, ya Fokotša Nako ya Tshepedišo ka 50%

Maitiro eMaruti Techlabs Akarerutsa Medicare Policy Kutengesa kuHealthPro Inishuwarenzi, Kucheka Kugadzira Nguva ne50%

மாருதி டெக்லேப்ஸ் ஹெல்த்ப்ரோ காப்பீட்டுக்கான மெடிகேர் பாலிசி விற்பனையை எவ்வாறு எளிதாக்கியது, செயலாக்க நேரத்தை 50% குறைத்தது

I-Maruti Techlabs Ekwenze Kwalula Ukuthengiswa Kwenqubomgomo Ye-Medicare Ye-HealthPro Insurance, Ukusika Isikhathi Sokucubungula ngo-50%

Stories, News & Mentions about PRESET

Categories

Trending Topics

Stories, News & Mentions about PRESET

Preset’s Stories on HackerNoon

Categories

Trending Topics