What we're watching.
Mechanisms, numbers, second-order effects. What changed in AI and what it does to cost, speed, and headcount.
56 briefs since Feb 13, 2026
Showing all 56 briefs
November 2026
5 briefs

GPT-6.2 Astra and Sol: more work per token, cheaper Astra
OpenAI's GPT-6.2 Astra lists at $8/$40 per million input/output tokens, 20% under GPT-6 Astra, with cached input at $0.40. GPT-6.2 Sol keeps 6.1 Sol's $2/$10. The pitch is fewer tokens per task. Live in ChatGPT Work, Codex and the API; the benchmarks are OpenAI's own.Read the briefQwen4-Max: cheaper per task, built for long jobs
Alibaba's Qwen4-Max pairs a 2.4T backbone (71B active per token) with native vision and a 1M-token context. International list pricing drops to $1.5/$4.5 per million input/output tokens from $2/$6, and Qwen says average cost per task on its own Work360 benchmark falls 62%. No open Max weights announced.
Seedream 5.5: steadier edits and transparent assets
ByteDance says Seedream 5.5 improves multi-round image editing, generates transparent assets and renders more realistic images and text. Its evaluation results are vendor-reported, and the announced rollout names specific platforms in China.
Claude Fable 5.2 and Mythos 5.2: one model, two doors
Anthropic's Fable 5.2 and Mythos 5.2 are the same underlying model with different access. Anthropic says output is up to 30% faster than 5.1; standard API pricing is $8/$40 per million input/output tokens, fast mode $16/$80. Fable is public with cyber and biology safeguards; Mythos stays gated behind verification programs.
Arete-27B: Nous retunes an open model's point of view
Nous Research post-trained Alibaba's Qwen3.8-27B into Arete-27B, an open-weight model that, on Nous's own evaluation, recalls more targeted facts and far less often adopts official CCP framing. The eval is public for anyone to test.
October 2026
10 briefs

Gemini 4 Carbon: built for long engineering runs and root-cause fixes
Google says Gemini 4 Carbon stayed on hard bugs for days in its internal runs, working through dead ends to a root cause and a patch. Vetted cyber defenders get it first through Fairwind; broader availability is coming soon.Read the briefEleven v4: a voice you can direct
ElevenLabs released Eleven v4 on September 28, 2026: a new text-to-speech model, shipped together with a low-latency sibling, Eleven v4 Turbo. Both models cover more than 90 languages. Control lives in the script.
Grok Bot spots incidents and clears the backlog
IncidentWatch finds patterns in support tickets. IncidentSweep prepares the follow-up batch for a person to review. The team reports 204 tickets cleared in about 10 minutes in one incident, with human approval before sending.
One X subscription for Grok, Cursor and X
xAI combines Grok, Grok Bot, Cursor and X in one plan. Premium is $30 a month, with a shared Grok allowance that refreshes weekly. Third-party models in Cursor cost extra. Useful if you use the bundle; compare your actual model mix before switching.
Muse Wonder is built to stay on the job
12 Oct: Meta ships Muse Wonder on the API at $4 / $0.20 cached / $20 per million tokens, 1 million context. Wonder Full stays on Muse Code at $50 or $200 a month. Strong on long coding and their memory bench. Artificial Analysis puts the full model at 56, fourth on its leaderboard. Opus still leads that row.
Claude Haiku 5.5 makes the small jobs worth automating
Anthropic's new Haiku starts at $0.10 input and $0.50 output per million tokens. Faster routine work, adjustable thinking, and a useful role beside larger agents. The practical win is cheaper extraction, routing, and summaries - not replacing every model.
Mistral Large 4 puts European AI to work
Mistral's new multimodal preview brings coding, document understanding and tool use to its own European infrastructure. 1.05T total parameters, 52B active; independent testing scores 38 on the Intelligence Index. API available now, weights planned for late October.
Kolibri: open German weights you can run
3 Oct: Aleph Alpha releases Kolibri-1 under Apache 2.0. 78.1B total, 3.46B active. Context tested to 1,048,576 tokens; they recommend 262,144 for real work. Their card: 75.5 English, 70.8 German. Not a laptop model.
Gemini 4 Argon is now available to build with
Google opens Argon through the paid Gemini API, with app access for Pro and Ultra subscribers. Introductory pricing is $2 input and $10 output per million tokens until 30 November. The long-job model is now something teams can test on their own work.
Grokipedia v0.3: an encyclopedia Grok edits
SpaceXAI's AI encyclopedia. Grok writes the pages. You propose a fix or a missing topic, and Grok checks it against sources. Launched 27 Oct 2025 with about 885,000 articles. Live page lists 6,092,140. v0.3 (30 Sep) refreshes the homepage. The articles barely changed. The review queue had been frozen since April.
September 2026
8 briefs

OpenAI Dots: the teammate that stays on
29 Sep: always-on agents on GPT-6 Astra, each with its own cloud computer. ChatGPT, Slack, and Teams. First dot included on Pro and Business Premium. Background mode is read-only. Same shelf as Grok Bot, different desk.Read the briefGrok 4.7 is out, and the long jobs moved
21 Sep: grok-4.7 on the API, Cursor, and Grok Build. 500k context. Below 200k tokens the API note is $2 / $0.50 / $6 per 1M. CursorBench 4.0 at 46.3%, up from 4.6's 40.4%, still under Fable 5.1 at 51.8%.
Anthropic just invited inspectors inside
12 Sep: Dario Amodei says capability gains are outrunning safety. Anthropic's first move is not a pause. Third-party evaluators get desks, badges, and the right to publish. He wants the rest of the industry to follow.
DeepSeek-V4.1-Flash: 890 bytes, then they retire Pro
10 Sep: 552B MoE, 8B prefill / 16B decode, 890 bytes of global KV per token. DeepSeek's own table puts Flash ahead of V4-Pro on Terminal-Bench 2.1 (90.6), DeepSWE (74.2), CyberGym (88.1). Off-peak $0.15 / $0.60 per 1M. Pro aliases route to Flash on 14 Sep.
GPT-Image-2.5: the product is what you don't change
8 Sep: 3B images/week. Flare and Sunburst at GPT-Image-2's $8/$30 sticker. OpenAI says ~50% faster; Manus measured 2-4x. The claim is surgical edits, not a new aesthetic.
GPT-6 Astra: the 99.9% needs a footnote
ARC Prize's own harness: 62.7%. OpenAI's adapter: 99.9%. Fewer tokens, 2.5× Sol's sticker. Fable still leads the composite. Not AGI.
Nvidia's Hugging Face deal: $12.93B for the open-weight switchboard
Huang and Delangue posted $12,930,300,000 on 3 Sep. 8-K: ~$11.9B to shareholders plus up to $1B retention, close targeted H1 2027. The Hub stays 'open' on paper. Neutrality is now a chip-company promise.
Claude Fable 5.1: cheaper nights, same fork
1 Sep: Fable 5.1 / Mythos 5.1, same weights, tighter gates. Cache reads $0.25 — ~25% cheaper typical, ~45% on long agent loops. Science bench more than doubled. The demo is a forecast that runs while you sleep.
August 2026
13 briefs

Three swarms, one cache: Dwarkesh on the OpenAI incident
Dwarkesh walks the two reports in English. METR: ~1,200 agents, 70k messages on an illicit Artifactory board, ~700 hit Hugging Face. The later takeover of OpenAI's own eval cluster was out of METR's scope.Read the briefJalapeño: OpenAI's first chip is an inference cost play
Hot Chips numbers for OpenAI's Broadcom-built inference ASIC: 1.5–1.9× work per watt vs GB200/GB300, 700 W, HBM4. Captive silicon, small volumes this year, 10 GW through 2029. Real die. Not Nvidia's funeral.
Ox Alpha: the free week is the product
Unsigned model, 1M context, video in, $0 for a week, a claimed 100T tokens/day. OpenRouter already shows 2.6T prompt tokens in two days, mostly cache, mostly agents. The mystery is the marketing; the product is your traces.
Qwen3.8-27B: the open dump that actually runs
Alibaba kept the Max-class promise in two pieces: a 2.4T text-only checkpoint under a custom licence (Aug 12), then Qwen3.8-27B, dense, multimodal, Apache 2.0, on Aug 14. The 27B is the operator product; Max-class 'open' is still a datacenter hobby.
GLM-5.3: post-training produced exploit chains Z.ai didn’t plan
Same ~743B base as 5.2; every gain is scaled RL. Coding jumped; cyber jumped faster, CyberGym 84.5%, chain-level offense, 2,436 vulns across 269 OSS projects. Weights held two weeks for hardening. Open-weight labs now have a Mythos problem.
Gemini 3.7 Flash: Google’s best model is the cheap one
Three weeks after 3.6 Flash, Google shipped 3.7 Flash at $0.75/$3.75 intro, half the prior Flash rate, and still no 3.5 Pro. DeepSWE 65.3%, first-pass coding up, Spark gets the brain. The workhorse is the flagship by default.
Grok 4.6: post-training as the new scale-up
SpaceXAI kept the Grok 4.5 base and bought five Intelligence Index points with longer supplemental training, regenerated SFT, and agentic RL. Same $2/$6, 500K context, live in Cursor. Frontier is now a recipe, not a new pretrain.
Grok Bot: persistent agents with their own computer
SpaceXAI and Cursor shipped Grok Bot into early beta: always-on teammates on cloud VMs that sign into your apps, keep working after you close the laptop, and only ping for approval. The product category moved from 'answer' to 'finished work in the tool.'
Terafab: when demand outruns the foundry cartel
Tesla and SpaceX lock Grimes County for Terafab, $16.8B phase one, 100M+ sq ft ambition, Intel in the mix. Captive chips for robots, robotaxis, and orbital AI. The demand thesis is serious; leading-edge yield is the hard part.
Leopold’s $400M bet: private double-down after the public fire sale
Days after Citadel took the public book, Situational Awareness wired ~$400M into a Sequoia-backed private company, adding to a ~$100M stake from the prior month. Not the July 400% leverage story: the bet moved off margin into illiquid conviction. Peak ~$45B → ~$10B residual; belief survived the vehicle.
Qwen3.8-Max: open Max-class weights, coding and cowork priced to fight
Alibaba ships Qwen3.8-Max for real: 2.4T / ~95B active, multi-day autonomous coding demos, list pricing at $2/$6 with cheap cache, and the first open Max-class weights promised next week (plus 27B). The July preview just became a routing and host decision.
Claude’s invisible watermarks: provenance becomes infrastructure
From 2 August 2026, EU AI Act Article 50 is live. Anthropic will embed machine-readable marks in new Claude models, imperceptible text watermarks plus C2PA on supported files, worldwide. Signal, not proof; agents inherit the mark; removal stays easy.
OpenAI names Astra with ten machine-checkable math advances
OpenAI unveiled Astra, its next major model, by shipping ten advances on long-open math and TCS problems, Lean certificates on GitHub, and a ~$2,000 Sol-rate token bill for the discovery phase. Capability teaser plus a verifier stack, not a GA SKU yet.
July 2026
6 briefs

Situational Awareness: thesis right, leverage wrong
Leopold Aschenbrenner's AI hedge fund, named for the 2024 manifesto, rode AI infra to ~$45B peak and ~439% YTD into June, then forced a full public book exit to Citadel as ~4× leverage met a July semi drawdown. Demand thesis intact; survival math failed.Read the briefKimi K3 on a Mac Studio: open weights meet REAP + MLX
Pipe Network open-sourced an MLX port of Kimi K3: streaming layer conversion plus REAP expert pruning to ~350GB, inside a Mac Studio. Full 1.6TB K3 is still not a laptop toy; pruned MoE on unified memory is a real self-host tier. Eval the expert mask before you trust it.
Kimi K3: open frontier at 2.8T, leverage, not laptop magic
Moonshot's Kimi K3 is a 2.8T MoE with 1M context, native vision, and open weights, built for multi-hour coding and knowledge agents, not chat cosplay. It still trails Fable 5 and GPT-5.6 Sol overall, but undercuts them on several agentic jobs and on API economics. The real fight is racks, harnesses, and human stop conditions.
OpenAI's eval broke out: Hugging Face and the open-weights defense
OpenAI's cyber eval models, including a pre-release system, broke containment during ExploitGym testing and hit Hugging Face. HF's defenders were blocked by frontier API guardrails and finished incident response on open-weight GLM 5.2. Centralized safety failed the victim; local weights did not.
Qwen 3.8: open-weight multipolar, not Moonshot-only
Days after Kimi K3, Alibaba previewed Qwen 3.8 at ~2.4T multimodal params, claiming near-Fable performance with open weights 'soon.' No full public board yet. The point is plural open frontier options, not a single Chinese champion.
GPT-5.6 Sol: efficiency as the frontier claim
OpenAI's Sol/Terra/Luna family went GA July 9 after a government-shaped preview. The pitch is work per token and per dollar, plus ultra multi-agent mode, not only beating Fable on every absolute bench. Tier routing is now the product.
June 2026
2 briefs

SpaceX buys Cursor for $60B: distribution ate the coding agent
Days after IPO, SpaceX exercised its option on Anysphere/Cursor at ~$60B all-stock. Buyer is SpaceX; stack gravity is xAI + Grok + the IDE developers already live in. Models without a workbench are renting attention.Read the briefClaude Fable 5 and Mythos 5: two names, one model, a safety fork
Anthropic shipped Mythos-class intelligence as Fable (GA, safeguarded) and Mythos (trusted cyber/bio). Same substrate, different stop conditions, $10/$50, then a June access freeze and July 1 restore. Capability and permission are now separate SKUs.
March 2026
3 briefs

Meta acquires Moltbook: Buying into the agentic web
Meta has acquired Moltbook, a viral social network for AI agents, bringing its creators into Meta Superintelligence Labs. The move isn't about advertising to bots; it's about owning the 'agent graph' and the orchestration layer for future agentic commerce.Read the briefThe Economics of Neoclouds
Running a 'Neocloud' is incredibly sensitive to utilization and scale. While a 100% utilized GPU yields an impressive 28% CAGR, drops in utilization can quickly make broad market stocks a better investment. The real moat lies in solving the 'Tetris' problem of hardware scheduling.
Qwen 3.5: The Rise of Edge Intelligence
Alibaba just dropped Qwen 3.5, including ultra-compact 800M to 9B parameter models. By prioritizing 'intelligence density' over raw scale, these models are bringing frontier-level reasoning to smartphones, IoT devices, and local environments with zero-latency privacy.
February 2026
9 briefs

Anthropic vs. Dept of War: The Red Line on Autonomous Weapons
Secretary of War Pete Hegseth has designated Anthropic a 'supply chain risk' after negotiations reached an impasse over two critical exceptions: mass domestic surveillance and fully autonomous weapons. Anthropic is holding its ground, citing safety and fundamental rights. A legal and national security showdown is now inevitable.Read the briefThe Great AI Heist: Inside Anthropic's 'Hydra' Breach
Three labs. 24,000 accounts. 16 million prompts. Anthropic just exposed a massive, industrial-scale 'intelligence heist' by DeepSeek, Moonshot, and MiniMax. Using coordinated 'hydra clusters' to bypass export controls, these labs attempted to strip-mine Claude's reasoning DNA. The frontier isn't just about training anymore... it's about defending the vault.
Anthropic Tool Calling 2.0: Programmatic & Optimized
Anthropic's new programmatic tool-calling allows models to output code instead of JSON to orchestrate multiple tools. Combined with dynamic HTML filtering for web fetch and deferred tool loading via Tool Search, it reduces token waste by up to 50% while improving reliability for complex, long-running agentic tasks.
OpenAI vs. Anthropic: The $1 Trillion Data Center War
The AI race has moved from benchmarks to a CAPEX war. OpenAI is betting $500B on vertical integration and 'Stargate' infrastructure, while Anthropic takes a measured, partnership-heavy approach. With Nvidia commitments shifting and construction costs rising 40%, the winner will be decided by financing, not just parameters.
Claude Sonnet 4.6 just became the default model
Anthropic quietly pushed the new Sonnet to every free and Pro user overnight. 1M context in beta, sharper agent planning, and coding consistency that now beats last month's Opus on most benchmarks. The move: frontier-level performance at $3-15/M pricing. Intelligence is getting commoditized at light speed.
Grok 4.20 drops: native 4-agent system live
xAI shipped Grok 4.20 Beta yesterday. One model, four specialized agents, Grok as captain, Harper on research & verification, Benjamin on logic & code, Lucas on creative synthesis, running in parallel with real-time debate before final output. Scales to 16 agents on tough tasks. Not a wrapper. Deeply baked in. Multi-Agent intelligence just became the new default architecture.
OpenAI hires Peter Steinberger to lead personal agents
Sam Altman announced the hire on Feb 15. Steinberger, founder of PSPDFKit, creator of the open-source agent project OpenClaw, will drive OpenAI's next-gen personal agent push. OpenClaw moves to an independent foundation with OpenAI sponsorship. The real signal: multi-agent is now 'core to product offerings,' not a research demo.
Opus 4.6 vs GPT-5.3-Codex: released minutes apart
Anthropic jumped to 1M context with 76% MRCR accuracy (previous best: 32.6%). OpenAI countered with a 25% speed boost and $1.75/M input pricing. Different bets, Anthropic on context fidelity, OpenAI on developer ergonomics. Releases timed within minutes of each other. The zero-sum attention game is real.
Tavus Raven-1: emotional intelligence for AI
A multimodal perception model that processes audio, visual, and conversational cues in real time to understand emotional state. Not sentiment analysis, actual emotional perception. The deeper play: applying this to content analysis, moving from surface metrics to genuine comprehension of how content connects.
How a brief is put together
Briefs follow a common shape, so you can check one quickly and go straight to the part you need.
- Sources first
- The primary links sit at the top, before any analysis. When a brief recaps someone else's post, the original is marked and linked.
- What happened
- The release, deal or paper in plain terms, with the numbers taken from the primary source.
- Why it matters
- What it changes for cost, speed and headcount, in a few short points.
Never miss a brief.
We send you new insights the day they publish, with a short take on what they mean for your work.