Search the archive

OPENAI'S NEW CPO

OpenAI has appointed Arvind KC as Chief People Officer.

BUILD DYNAMIC AGENTIC WORKFLOWS IN OPAL

Google has introduced agentic intelligence to Opal workflows, enhancing flexibility and interactivity.

ANTHROPIC OFFERS STAFF $6B SHARE SALE AT STAGGERING $350B VALUATION

The transaction could total between $5 billion and $6 billion.

WE ARE CHANGING OUR DEVELOPER PRODUCTIVITY EXPERIMENT DESIGN

METR is changing its Developer Productivity Experiment design as a significant number of developers are choosing not to participate in the study as they do not wish to work without AI.

See how it works →

ANTHROPIC REPORTED LARGE-SCALE DISTILLATION ATTEMPTS

Anthropic accused DeepSeek, Moonshot AI, and MiniMax of creating over 24,000 fake accounts to generate roughly 16 million Claude interactions aimed at replicating its agentic reasoning, tool use, and coding capabilities.

OPENAI PREPARES NEW CHATGPT PRO LITE TIER PRICED AT $100 MONTHLY

OpenAI plans to offer a new subscription tier called ChatGPT Pro Lite. At $100 per month, the new plan could serve users who hit Plus rate limits regularly but do not need the unlimited access that Pro delivers. It could be a way to accommodate Codex usage as well. No release date has been confirmed yet, and the exact feature set remains unclear.

PERPLEXITY TESTS MESSAGES INTEGRATION AND USAGE CREDITS

Perplexity is testing a Messages connector for its Comet browser, enhancing local integration with Apple's Messages app. This appeals to macOS power users by merging communication history with web data. A new "Usage and Credits" feature allows users to buy additional credits, addressing backlash from reduced Pro plan limits.

GPT-5 CODEX RAN A 25-HOUR CODING SPRINT

OpenAI described a long-horizon stress test where GPT-5.3-Codex, given a blank repository and full access, built a design tool over ~25 hours using ~13M tokens and producing ~30k lines of code.

SWE-BENCH VERIFIED STOPPED BEING A FRONTIER CODING METRIC

SWE-bench Verified no longer reliably measures frontier coding progress because many tasks have flawed tests that rejected correct fixes and because benchmark contamination has made training exposure a significant factor in scores.

MATHEMATICS IN THE LIBRARY OF BABEL

Daniel Litt argues that current limitations are real but likely temporary, not fundamental blockers. Litt now expects AI to autonomously resolve mildly interesting open conjectures as early as late 2026 and suggests AI's long-term impact on mathematics may exceed that of the computer. Even in a future where machines can generate vast bodies of correct math, he argues human mathematicians will still matter because the core motivation remains understanding, not just producing results.

credits/month

AWS LAUNCHES STRANDS LABS TO GIVE DEVELOPERS A SANDBOX FOR AUTONOMOUS AI

AWS Strands Labs is a new GitHub organization aimed at helping developers explore and experiment with cutting-edge AI techniques. It is currently available for three specific projects: robots, robots sim, and AI functions. AWS considers these areas the ones that most clearly reveal what changes when software becomes agentic. Strands Labs gives AWS and the broader community a dedicated space to experiment boldly.

THE FIRST FULLY GENERAL COMPUTER ACTION MODEL

FDM-1, a foundation model for computer use, efficiently trains on 11 million hours of video to autonomously handle tasks like CAD, driving, and fuzzing. Unlike traditional models that rely on expensive, limited annotation, FDM-1 uses a video encoder that compresses 2 hours of video into 1M tokens and predicts actions with inverse dynamics for scalable training. With enhanced context processing, FDM-1 showcases significant advancements in executing long-horizon tasks and exploring complex state trees in GUI environments.

SELF-AWARE REASONING EFFICIENCY

SAGE introduces a self-aware sampling approach that enables large reasoning models to decide when to stop thinking, reducing unnecessary computation. Its companion method, SAGE-RL, distills these efficient reasoning patterns into standard pass@1 inference, reporting accuracy improvement with fewer tokens across six challenging math benchmarks.

MACHINEAUTH

MachineAuth provides secure OAuth 2.0 client credentials flow for AI agents, enabling machine-to-machine authentication. Key features include JWT token generation, agent management, and scope-based permissions, all with simple JSON file storage and no databases required.

TAALAS HC1: ABSURDLY FAST, PER-USER INFERENCE AT 17,000 TOKENS/SECOND

Taalas has been quietly building a model-on-silicon chip that bakes an LLM directly into hardware to deliver absurdly fast, per-user inference. The HC1 chip runs Meta's Llama 3.1 8B at extreme speed by effectively hardwiring the model into silicon. It delivers inference at around 17,000 tokens per second per user. This is just the first version - the next iterations are already geared toward much better fidelity.

OPENAI LANDS MULTIYEAR DEALS WITH CONSULTING GIANTS IN ENTERPRISE PUSH

OpenAI is entering into multiyear partnerships with consulting firms Accenture, Boston Consulting Group, Capgemini, and McKinsey & Co. to help deploy its enterprise platform, Frontier. Frontier is an intelligence layer that stitches together disparate systems and data within an organization to make it easier for companies to manage, deploy, and build AI agents. OpenAI decided to partner with the consulting firms because they have existing relationships with enterprises and deep knowledge about how those businesses operate. They will help OpenAI roll out its technology at scale.

FULLY REMOTE)

As the first engineer on TLDR's new Applied AI team, you'll build AI agents and composable Claude Skills to let non-technical teammates create their own AI workflows. Learn more [21].

IBM STOCK DIVES AFTER ANTHROPIC POINTS OUT AI CAN REWRITE COBOL FAST

Anthropic claims that Claude Code tools can accelerate the refactoring of apps written in COBOL.

DEEPSEEK V4: RUMORS VS REALITY FOR THE NEXT BIG CODING MODEL

DeepSeek V4, a highly anticipated open source model with a 1M+ context based on Engram architecture, will likely be released before the end of the month.

ASML ANNOUNCES BREAKTHROUGH IN EUV LIGHT SOURCE TO BOOST CHIP OUTPUT

The development could enable chipmakers to produce up to 50% more semiconductors by the end of the decade.

Register now

MICROSOFT DEVELOPS COPILOT ADVISORS TO DEBATE ON ANY TOPIC

Microsoft is developing Copilot Advisors for debating any topic using AI personas like legal or finance experts. Users choose two agents who argue with distinct voices, possibly with animated portraits. This feature aims to enhance individual analysis and decision-making by presenting strong cases from both sides.

CLAUDE CODE SECURITY RESEARCH PREVIEW

Anthropic introduced Claude Code Security in a limited research preview, enabling automated scans of codebases to surface context-dependent vulnerabilities and propose targeted patches for human review.

ANTHROPIC RESEARCH: MEASURING AI AGENT AUTONOMY IN PRACTICE

Anthropic announced the release of its improved Claude LLM, focusing on enhanced safety and performance. The update includes better handling of complex queries and increased integration capabilities for businesses. This aims to provide improved utility in various professional applications.

THE FILE SYSTEM IS THE NEW DATABASE: HOW I BUILT A PERSONAL OS FOR AI AGENTS

Personal Brain OS is a file-based personal operating system that lives inside a Git repository designed to stop agents from forgetting important context information. It contains everything the AI assistant needs to recreate users' voice, brand, goals, content pipeline, research, and more. This article discusses how Personal Brain OS was built, why the architectural decisions matter, and the lessons the developer learned the hard way. The full architecture is available for anyone to clone and make their own.

WHY DEVELOPERS KEEP CHOOSING CLAUDE OVER EVERY OTHER AI

Developers keep choosing Claude Code because the alternatives keep failing them in the same way. Generating correct code is a major part of the job, but it isn't everything. Agents still need to edit files without corrupting the surrounding code, read the right files before making changes, complete multistep tasks without losing threads halfway through, know when to ask questions, and more, something no other coding agent does consistently. Claude Code stays on track more than other tools, making it the preferred choice.

FREE YEAR → framer.link/TLDRAI

Startups move fast, and design should too. With Framer,

early-stage companies can launch a beautiful, production-ready site in hours. Now you can get one year of Framer Pro free [11] and join thousands of founders already building on Framer. APPLY TO CLAIM YOUR

OPENAI FIRST PROOF SUBMISSIONS

OpenAI detailed its end-to-end proof attempts on all ten First Proof research problems, a domain-specific challenge designed to test whether AI systems can produce fully checkable mathematical arguments. The post outlined evaluation feedback from field experts, revisions to earlier correctness claims, and included a preprint with all proofs plus an appendix documenting prompt patterns used to simulate iterative human–model collaboration.

THE CODING AGENT IS DEAD

The current generation of coding agents isn't the future. The newest models can be powerful with nearly any tool thrown at them, with bash often being enough. Codebase organization is now the bottleneck. Amp plans to kill the Amp editor extensions for VS Code and Cursor on March 5. It will continue to work on the Amp CLI, which is flexible, light, easy to change, and can be run anywhere at anytime.

ALPHAEVOLVE DISCOVERS NEW MULTI-AGENT LEARNING ALGORITHMS

Using an evolutionary coding agent powered by large language models, the work automatically generated new algorithms for imperfect-information games across CFR and PSRO paradigms. The discovered VAD-CFR and SHOR-PSRO variants introduced adaptive discounting and hybrid optimistic meta-solvers, showing improved empirical convergence over strong baselines.

LEVERAGING OPENCLAW AS A WEB DEVELOPER

Web developers can integrate OpenClaw to automate repetitive tasks like scaffolding routes, generating tests, and updating documentation directly from natural‑language prompts. Developers with domain context can teach OpenClaw project‑specific patterns that improve code quality and consistency without manual intervention.

CHATGPT ADS NOW HIT ON YOUR VERY FIRST MESSAGE

OpenAI is now testing ads in ChatGPT for Free and Go users in the US, with companies like Expedia and Qualcomm appearing after a user's first message. The beta charges $60 CPM with a $200,000 minimum investment, only including large-budget brands, and excludes paid subscribers.

considerations show up. Download the security flashcard

WHY IS CLAUDE AN ELECTRON APP?

Claude is an Electron app because developer familiarity and simpler maintainability across multiple platforms are worth the tradeoffs.

WHICH WEB FRAMEWORKS ARE MOST TOKEN-EFFICIENT FOR AI AGENTS?

Minimal API web frameworks are far quicker and more cost-effective for agents to work with.

REINFORCEMENT LEARNING FOR LLMS

A look at the surprisingly simple underlying principles of reinforcement learning.

APPLE'S UPCOMING AI SMART GLASSES ARE STARTING TO SOUND A LOT MORE EXCITING

Apple is accelerating development of AI smart glasses featuring two integrated cameras, aiming to rival Meta Ray-Bans.

GEMINI 3.1 PRO

Google released Gemini 3.1 Pro as the upgraded core model behind recent Gemini 3 “Deep Think” improvements, and began rolling it out to the Gemini API/AI Studio, Vertex AI, Android Studio, the Gemini app, and NotebookLM. The post highlighted a verified 77.1% score on ARC-AGI-2, more than doubling Gemini 3 Pro's result.

GOOGLE TEST NOTEBOOKLM INTEGRATION FOR OPAL WORKFLOWS

Google tests NotebookLM integration within Opal workflows, enhancing data extraction and automation. This integration aims to streamline processes and improve workflow efficiency for users.

ARC-AGI-3 UPDATE

ARC-AGI-3 is an Interactive Reasoning Benchmark designed to measure an AI Agent's ability to generalize in novel, unseen environments. Opus 4.6 demonstrates better reasoning and use of memory than Gemini 3.1 Pro and solves more levels. Current models may be able to solve ARC-AGI-3 given access to a harness with a simple memory. Memory scaffolds are likely enough for pseudo-continual learning to push us to some self-improvement or research-agent threshold within the next 2 years.

AI #156 PART 1: THEY DO MEAN THE EFFECT ON JOBS