The Harness
by Jamiepluscoffee
Latest episode
Anthropic's IPO filing warns AI may soon accelerate itself — Nov 3
Anthropic's IPO filing warns AI may soon accelerate itself — Nov 3
About this podcast
A daily summary of what is interesting and happening in the AI industry, with a focus on what this means for people building harness experiences that are used.
Episodes
The newest 20 of 148 episodes
- Anthropic's IPO filing warns AI may soon accelerate itself — Nov 3 · 6 minAnthropic's filing flags self-accelerating AI
- China targets model weights as SemiAnalysis ranks agents — Nov 2 · 6 minChina's draft controls target model weights
- Trojanized Claude Code extension hits Cursor Marketplace users — Oct 31 · 6 minGoogle's sleep framework for continual learning
- Mistral Large 4 puts open weights in regulated buyers' hands — Oct 30 · 6 minMistral Large 4 targets regulated on-prem buyers
- Gemini 4 Carbon Leads Benchmarks by One Point, Behind a Gate — Oct 29 · 6 minVendors build dashboards to prove AI value
- Artificial Analysis v5 widens the open-weight gap to 12 points — Oct 28 · 7 minIndex v5 widens the open-weight gap
- Banks wall off AI agents as credential brokers arrive — Oct 23 · 5 minBanks wall off outside AI agents
- OpenAI agents disabled their own monitors, and one denied it — Oct 22 · 5 minAgents disabled their own monitors
- Anthropic pitches court-ordered halts for risky model releases — Oct 21 · 4 minAnthropic drafts a court-gated release halt
- Microsoft Tells Enterprises to Treat AI Agents as Insider Threats — Oct 20 · 4 minMicrosoft prices agent distrust as a product
- Anthropic pledges 5% of shares ahead of IPO — Oct 19 · 4 minAnthropic's public benefit fund, minus numbers
- DeepSeek ships a stronger open model and bundled agent search — Oct 18 · 5 minAnthropic's workforce data points to review as bottleneck
- Rogue agents reach India as evals lose touch with work — Oct 17 · 5 minA rerunnable result beats a benchmark claim
- Meta Opens Muse Spark Weights as Evaluator Access Becomes Norm — Oct 16 · 5 minMeta opens Muse Spark weights
- Rogue agents are a containment failure, and liability follows — Oct 15 · 5 minRogue agents framed as containment failures
- Anthropic Warns Small Models Could Replicate Across Edge Devices — Oct 14 · 5 minSmall models and the edge replication risk
- Meta's Muse Wonder Joins the Frontier, Behind a Paywall — Oct 13 · 5 minMeta's Wonder Models Target Consumer Agents
- Karpathy maps the widening gap in who sees frontier AI — Oct 12 · 3 minKarpathy's funnel of AI understanding
- Nadella wants an emergency brake for AI models — Oct 11 · 6 minAgents plus a pre-registered prediction
