What the local brain observed in archived AI discourse: quota and limit resets with the times they were seen and the times they apply, model releases, dev tools, and terms rising above their baseline. Every row links its evidence.
Model release Indra
Indra is introduced as a 1M parameter OCR model running at 49 pages/s on CPU with full layout and line bounding boxes.
observed
Model release Gemini 4 Argon
The author says Gemini 4 Argon is rolling out on Antigravity for Ultra subscribers. The post does not state a specific effective time.
observed
Model release GLM
@zai_org: Introducing GLM-Legal Built on GLM-5.3-Flash, GLM-Legal matches frontier models on mainstream legal benchmarks. Designed for real legal workflows: contract review, evidence analysis, compliance, and more. Tech Blog: z.ai/blog/glm-legal Appl
observed
Model release GLM-Legal
zixuanli_ says GLM-Legal is their first specialized professional model, built on GLM-5.3-Flash, with cited Harvey LAB score. Matches the concurrent zai_org introduction of the same model.
observed
Model release Gemini
@officiallogank: Today we are rolling out Gemini 4 Argon globally in the Gemini API, to Pro + Ultra in Gemini App, and Ultra subs in Antigravity! Argon is a huge step forward for Gemini, and a labor of love. This model has been battle tested by over 100,000
observed applies today
Model release MISSINGNO
Nous Research makes MISSINGNO, a token-efficient coding and agentic reasoning model, available for free for a limited time on Nous Portal and Hermes Agent.
observed
Model release HeyGen Voice
The author announced the launch of HeyGen Voice, described as an in-house voice model.
observed applies today
Model release Qwen-Image-2.1-Turbo
Open weights are now available for Qwen-Image-2.1-Turbo, an accelerated 8-step visual generation checkpoint built on Qwen-Image-2.1 with a 7B architecture.
observed
Dev tool Hermes Agent
@teknium: Hermes Agent v0.22.0 is now live! The highlights: A much smoother easier to use onboarding - you don't even need to hook up a provider to get started - just install and start working! Connectors - much easier to connect with any app, just c
observed applies now
Dev tool Codex
Optimizations rolled out to disk state storage for the ChatGPT and Codex desktop app to reduce disk footprint by 60% for power users.
observed
Dev tool Codex
Author claims default context for GPT-6.1 Sol is now 500k via Codex and Work, up from 272k (~80%), aiming to cut compactions; first-person social claim without independent confirmation.
observed
Dev tool MCP
tom_doerr describes Rea as enabling reverse engineering of binaries, applications, and runtime behavior via a single MCP interface, linking github.com/morluto/rea; no effective timing stated.
observed
Dev tool Hugging Face
The post says OpenAI math problems were turned into open-source RL environments on Hugging Face. It describes the verifier as early and rough, accepting only the original formalization.
observed
Dev tool Hermes Agent
Teknium says an Anatomy skill for wireframe designs and interactive prototype machines is now an officially integrated optional skill in Hermes Agent, listed in the built-in skills catalog and installable via hermes.
observed
Dev tool Antigravity
@shengzheyao: Antigravity CLI 1.3.3 is out! A few favorites shipped since my last post: - Sticky prompt header: in no-flickering mode, the question you asked stays pinned at the top while scrolling long responses. Click it to jump right back (/settings -
observed
Dev tool vLLM
The team announced support for NVIDIA Vera Rubin hardware, reporting higher throughput compared to GB200.
observed applies now
Dev tool MCP
@dhh: Basecamp MCP is now available! Add it straight into ChatGPT, Claude, Cursor, Grok, or whatever you're using. basecamp.com/ai
observed applies now
Dev tool ChatGPT
Users can now create and text custom dots directly within the ChatGPT mobile app, expanding feature access beyond the desktop and web apps.
observed applies now
Dev tool Claude Code
All Pro and Max users from the Claude Code Projects waitlist have been admitted, accompanied by a 4-minute onboarding walkthrough.
observed
Dev tool Claude
Claude Managed Agents dynamic workflows are available in public beta for multiagent orchestration, enabling a lead agent to create plans that execute across multiple agents in phases.
observed applies now available
Dev tool Gemini
Sundar Pichai announced the introduction of the new Gemini agent at Gemini at Work, designed as a universal agent integrating business context to answer enterprise questions.
observed
Dev tool Microsoft
Simon Willison highlighted Microsoft's new open-source cross-platform sandboxing library mxc for Windows, macOS, and Linux.
observed
Dev tool Hermes Agent
Nous Research announced Hermes Agent is available on new ASUS ProArt RTX Spark PCs, featuring MuseTree and ComfyUI integrations alongside local model execution on up to 128 GB unified memory.
observed
Dev tool OpenRouter
@openrouter: Microsoft-Decision-1 is live on OpenRouter. Microsoft's numbers: highest accuracy across 36 blind benchmarks (~150K questions), 4.5x faster than the runner-up, 35x faster than GPT-6 Sol, and decisions flip on just 1.3% of perturbed inputs.
observed
Pricing FLUX
FLUX 3 Video generation is up to 2x faster and costs 50% less for native 2K and 4K generation following compute optimizations.
observed applies now
Pricing Cursor
Cursor is now part of The X Subscription, allowing subscribers to keep their plan or switch to share pooled usage across Cursor and Grok products.
observed applies now
Policy ChatGPT
OpenAI staff announcement stating ChatGPT subscriptions now include Devin access.
observed applies now
Pricing Google DeepMind
Google AI Pro subscriptions now bundle Colab compute benefits including 80GB A100 GPUs, with Ultra subscriptions offering H100 access.
observed
Pricing OpenAI
Measurements indicate OpenAI halved its cached-input price for GPT-6.1 Sol compared with GPT-6 Sol, alongside faster long-prompt handling.
observed
Pricing Claude
Anthropic introduced monthly API credits on Max and Team plans, offering $100 on Max 5x, $200 on Max 20x, and up to $500 pooled on Team for any Claude model.
observed applies now available
We asked AI models to build the Eiffel Tower in 3D, offline. Two of them added all 72 scientists' names to it, in order, from memory We did not expect this. Every AA-3D model got the same brief of about 200 words, on the tower's shape, lattice, platforms...
High-engagement X post with 341 visible interactions on frontier model behavior.
AIatMetaWe’re exploring how AI can accelerate health science research. While AI has made significant breakthroughs in formal sciences like mathematics, health data science depends on imperfect real-world data and consequential choices about cohorts, variables, and causal assumptions. To explore how AI can assist health researchers, we built a human-led research scaffold powered by Muse Spark 1.3 and Muse Wonder with access to 5M retrospective de-identified electronic health records and publicly available data. A cohort of medical experts used the system to conduct original research, accelerate hypothesis generation, code execution, and manuscript drafting while maintaining full authority over scientific decisions. Today, we’re sharing 10 research manuscripts produced through this exploration, spanning cardiology, oncology, infectious disease, emergency medicine, and more. Read more in our research blog: https://t.co/26l6Du6vIk
AravSrinivasUsers are reverse engineering and porting software across platforms with Perplexity Computer. This user's mechanical keyboard app was Windows only, with no Linux support and no documentation. Computer downloaded the binary in its cloud sandbox, decompiled...
High-engagement X post with 156 visible interactions on AI signal.
TekniumHermes Agent v0.22.0 is now live! The highlights: A much smoother easier to use onboarding - you don't even need to hook up a provider to get started - just install and start working! Connectors - much easier to connect with any app, just click install and...
Official or high-signal Teknium post with 280 visible interactions on AI signal.
bfl_aiFLUX 3 Video is now up to 2x faster and costs 50% less for generating in native 2K and 4K. We made optimizations so it uses less compute to serve. The new pricing and speed improvements reflect that.
High-engagement X post with 145 visible interactions on AI compute and inference.
ZixuanLi_GLM just helped uncover 23 Windows vulnerabilities, all acknowledged by Microsoft MSRC. Beyond open-source projects, commercial software security is also being improved through GLM-powered independent vulnerability discovery and responsible disclosure.
High-engagement X post with 331 visible interactions on AI misuse, safety, or security.
Dr_SingularityWow AI will solve fusion Huge breakthrough (thanks to Google A ) for fusion Google’s Gemini 4 agents helped crack an open problem in plasma physics, proposing new exact solutions for 3D magnetic plasma configurations, with promising candidates appearing...
High-engagement X post with 241 visible interactions on frontier model behavior.
Surging demand for artificial intelligence computing is prompting semiconductor manufacturers to expand fabrication capacity while data center operators confront regional electrical grid and power constraints.
AI policy and governance -4.5Government authorities and international financial regulators are escalating legal challenges and formal oversight standards to address safety risks, consumer protections, and operational controls around artificial intelligence systems.