What mattered
The week in AI, cut for builders.
Builder conversations · weekly
Weekly Claw
Sharp conversations with the people building AI: agents, devtools, and the future of work.
Fridays · 4 PM Eastern
Featured latest episode · October 30, 2026
Wall Street became AI’s landlord. And the numbers started lying.
Recurring formats
The week decides
No rigid podcast template. Every Friday runs the conversations, arguments, and demonstrations that the week earned.
Hot takes
The arguments in DMs, now on record.
Builder demo
Something real, shipped and questioned live.
Tool fight
Two tools, one task, one verdict.
Agent autopsy
What broke, why, and what it teaches.
Guest builder
The people shaping software’s next move.
Latest Episodes
W36
Video + slides
Wall Street became AI’s landlord. And the numbers started lying.
Wall Street became AI's landlord — and the numbers started lying. Banks securitized the chips ($63B SPV, ~$40B GPU lease) and Meta filed ~$95B of equity paper hours…
W35
Video + slides
The open lane got the best model of the week. Then it lost every comfortable place to run it.
The open lane had its best week on paper and its worst week in practice. DeepSeek shipped MIT weights that beat GPT-5.6 Sol on Terminal-Bench 4.0, MiniMax revealed…
W34
Video + slides
Local AI, Voice Models & Building Reliable Agents | Weekly Claw 34
A 74B model fit into 24GB with training data attached, a 600B model opened its weights, a voice model topped the speech index — and a boardroom recorder, 406 scientists and a $16.8B quarter landed in the same week.
W33
Video + slides
Ultrafast Agents, Mistral Large 4 & OpenAI’s Math Release | WeeklyClaw #33
Open-weight previews put scale claims beside receipts: Mistral Large 4 and Reflection Beam, OpenAI’s math manuscripts, Windows execution containers, agent reliability data, and the consumer-agent debate.
W32
Video + slides
GPT-6.1 Sol, Gemini 4 Argon & NVIDIA’s Agent Safety Debate | WeeklyClaw #32
GPT-6.1 Sol capability and cost, Pro allowances, dots and Decisions API, restricted Gemini 4 Argon, OpenClaw enterprise, and NVIDIA’s agent safety debate.
W31
Video + slides
Opus 5.5's Price War, GPT-6 Sol & Luna, Grok 4.7 & Financial AI | Weekly Claw #31
Seven days after the industry's loudest safety week, its answer was a price war: Anthropic shipped Opus 5.5 at Fable-tier quality for 40% less, OpenAI answered within…
W30
Video + slides
Narrow Models Win, Typed Decisions, Union Alpha, Signal From Outside
Andy Lauppe and Henry Mascot cover narrow models that decide with typed probabilities, Union Alpha on OpenRouter, and signal from outside the lab.
W29
Video + slides
OpenAI’s Maths Claim, Meta Muse Hands-On, AI Harnesses, The Safety Debate
Andy Lauppe and Henry Mascot examine OpenAI’s claimed mathematics result, try Meta Muse, compare agent harnesses and debate AI safety.
W28
Video + slides
GPT-6 Astra arrives, OpenClaw 2.0, Jensen Huang on AI and work, NYC’s classroom AI pause
Andy Lauppe and Henry Mascot discuss GPT-6 Astra, OpenClaw 2.0, the week’s model launches, Jensen Huang’s comments on AI and work, and classroom AI policies.
W27
Video + slides
The agent owns the loop.
This week was the consolidation race. Nvidia moved to buy the open-weights commons outright — $12.9B reported for Hugging Face. Zhipu unmasked its viral stealth model…
W26
Video + slides
The operating layer became the company.
The operating layer beneath the model became the company. Inherent post-trained a 27B supervisor to direct GPT-5.5 Codex like a scientist; Cerebras CS-4 and Inco…
W25
Video + slides
The OS got built this week.
The model is no longer the product — the operating layer is. This week DeepSeek shipped weights, harness, and Open Responses + Anthropic Messages API dialect together…
W24
Video + slides
The harness ate the model.
Capability barely moved this week. The control plane did. The most consequential AI receipts of the past seven days are no longer new frontier weights — they are open…
W23
Video + slides
The envelope, not the engine.
Capability barely moved this week; everything around capability did. The economics moved — OpenAI cut Luna's price 80% three weeks after launch. The openness moved —…
W22
Video + slides
The Sandbox Failed
Welcome back to Weekly Claw. Today's episode has a shape to it, so let's name it up front: risk, then control, then architecture, then a new place to work, then who…
W21
Video + slides
The model is portable. The workflow is the moat.
Last week was the model price shock. This week is the ownership shock. The story is not just “another big model launched.” The story is that the center of gravity is…
W20
Video + slides
The cost of intelligence collapsed
GPT‑5.6, Grok 4.5, agentic ransomware, and a Fable-vs-GPT Tool Fight.
W19Main + changelog
GPT‑5.6 Sol and malicious agent skills
Agent security, ClawExplorer events, and the future of weird software.
W18Changelog / DevX
Community signal and weekly summary
OpenClaw release notes, developer-experience signal, community updates, and the week’s close.
W17Changelog / DevX
Release proof loops
Maintainer signal, release evidence, and a developer-experience readout grounded in what shipped.
W15Main + changelog
Enterprise agents and token economics
Claw Chain, enterprise governance, robot arms, and the first meetups.
Supported by
Come on the show
Built something weird? Prove it live.
Pitch a demo, nominate a tool fight, bring the agent failure everyone can learn from, or follow the show where the clips and weekly notes land.