OpenAI Introduces New $100 Pro Tier

#1 𝕏

OpenAI launched a new $100/month ChatGPT Pro tier offering 5× the Codex usage of Plus (boosted to 10× through May 31) for longer, high-effort sessions. It also includes all existing Pro features—exclusive Pro model access and unlimited Instant and Thinking models.

#2 𝕏

Google DeepMind launched Gemma 4, a lineup of 7B–196B-parameter foundation models with up to 100K-token contexts and multimodal capabilities. Developers can now access open-source weights, code samples, and tutorials via Vertex AI and GitHub to jumpstart building AI apps.

Also covered by: @Jeff Dean

#3 𝕏

Claude launched the Advisor Strategy beta on the Claude Platform, introducing a new guided AI advisory feature.

#4 𝕏

Claude launched Claude Cowork for all paid plans and added enterprise features—role-based access controls, group spend limits, usage analytics, and expanded OpenTelemetry—for org-wide deployment.

#5 𝕏

DeepLearning.AI: An accidental leak exposed over 500,000 lines of Anthropic’s Claude agent code, revealing its modular tools, subagent swarms, and layered memory management.

#6 𝕏

Google Research introduced ConvApparel, a human-AI conversation dataset paired with an evaluation framework to quantify and bridge the “realism gap” in LLM-based user simulators, boosting the training of more robust conversational agents.

#7 𝕏

Rowan Cheung: Meta launched TRIBE v2, a foundation model trained on 1,000+ hours of fMRI data from 720 people that predicts which brain regions light up, how strongly, and in what order from video, audio, or text—outperforming real scans.

#8 in

Dharmesh Shah launched jsondata.com, a free AI-powered online tool for viewing, filtering, compressing, and manipulating JSON data in a nested interface.

#9 𝕏

Guillermo Rauch unveils Agentic Infrastructure, a paradigm that treats cloud infrastructure as autonomous coding agents (e.g., Claude Code, Vercel) to automate deployment and operations.

#10 𝕏

Lenny Rachitsky spotlights Simon Willison’s “lethal trifecta”: AI agents with private data access, untrusted content intake, and exfiltration capability pose a massive security risk that only dropping one of these legs can solve.

#11 𝕏

Cognition warns that 92% of COBOL developers will retire in the next four years and 68% of enterprise COBOL modernization projects are failing, and outlines how software agents can streamline and accelerate COBOL modernization at Fortune 500 companies.

#12 𝕏

LlamaIndex 🦙 launched LlamaParse and LiteParse Agent Skills, giving AI agents access to layout, tables, images and structured context in PDFs and other unstructured docs for more reliable knowledge extraction and automation.

#13 𝕏

Jeff Dean asked Gemini to analyze all billboards listed on 101ads.org and generate a report categorizing each company by industry.

#14 𝕏

Philipp Schmid shared five essential principles from his talk on why senior engineers struggle with AI agents: treating text as state, handing over control, viewing errors as inputs, shifting from unit tests to evals, and designing evolving agents instead of static APIs.

#15 𝕏

Andrew Ng unveiled a new short course, “Efficient Inference with SGLang: Text and Image Generation,” co-built with LMSys and RadixArk and taught by Richard Chen, teaching how to use SGLang’s open-source caching framework to slash redundant LLM costs by processing shared promp...

#16 𝕏

Santiago: They’ve built a completely new Large Memory Models architecture that mimics human memory instead of using RAG or vector search. The founders—authors of 160+ Nature and ICLR papers—even closed their Harvard lab to focus on it.

#17 𝕏

clem 🤗 argues the eval likely just ran Semgrep or CodeQL to spot bugs, so it isn’t an apples-to-apples comparison, and hopes open-source models will match closed-lab capabilities.

#18 𝕏

Josh Woodward rolled out full access to Lyria 3 on GeminiApp, letting users generate up to 5 free full-length (~3 min) tracks per day and unlimited 30-second clips, with paid upgrades for extra music, images, and video.

#19 𝕏

Santiago introduces an AI agent that auto-joins meetings, reads Slack and hooks into your tools to build a unified team context—fueling automated workflows, seamless new-hire onboarding, and real-time risk detection.

#20 𝕏

Cursor can now attach demos and screenshots of its work to the PRs it opens, letting your team review cloud agent–generated artifacts directly in GitHub.

#21 𝕏

Alexandr Wang rolled out a full revamp of Meta AI’s product stack alongside the Muse launch—introducing a brand-new app and a suite of freshly launched features.

#22 in

Peter Yang reports that Silicon Valley AI tools—from Cursor’s Composer 2 on Moonshot’s Kimi K2.5 to Cognition’s SWE-1.6 fine-tuned on Zhipu’s GLM and Airbnb’s reliance on Alibaba’s Qwen—are all powered by Chinese open-source models. He highlights Zhipu’s new GLM-5.

#23 𝕏

Andrej Karpathy suggests OpenClaw’s breakout moment came because it was the first time many non-technical users—who until then equated AI with the ChatGPT website—actually got hands-on with advanced agentic models.

#24 𝕏

Andrej Karpathy warns that many people’s outdated views—shaped by last year’s free ChatGPT tier and viral glitch clips—miss the massive leaps in capability of today’s agentic AI.

#25 𝕏

Harrison Chase notes that community middleware—like “langchain-task-steering”—is popping up for customizing agents and deepagents, and invites anyone with middleware to contribute by reaching out to Sydney.