Summary
Thin evening slot dominated by off-topic political and Elon-meme reposts from @brivael, with only two items worth surfacing. The standout is Google Research announcing ERA (Empirical Research Assistance), an agentic coding system published in Nature and pitched as enabling scientific breakthroughs, which lines up with the broader auto-research thread the wiki has been tracking. Scobleizer plugs LobeHub as an "agent workspace" framing rather than a chatbot, an interesting product-shape signal even though the post is promo-flavored. The rest, French politics, taxpayer-fraud videos, "100% AI companies will demolish" Musk clips, is signal-free for AI research purposes.
Posts
- Google publishes ERA (Empirical Research Assistance) in Nature (@GoogleResearch). Agentic coding system from Google Research, framed as unlocking "previously impossible" scientific breakthroughs, presented at I/O. Sits squarely in the auto-research / agent-driven science thread already covered by autoresearchclaw and the auto-research roadmap. Worth a wiki summary once the Nature paper is in hand.
- LobeHub framed as an "AI workspace" / Chief Agent Operator (@Scobleizer · lobehub.com). Product positions itself as a long-horizon agent orchestrator that hires, schedules, and reports on a team of agents (312K skills, 58K MCP servers claimed). Signal here is the shift in product framing from chatbot-UI to operator-of-agents, not the tool itself. Promo-leaning but the shape matters.
- "The killer product is still the CPU" (@magicsilicon). One-line take pushing back on GPU-only narratives. Provocative but unsupported, log it and move on.
- Off-topic noise cluster (cluster of 17) (@brivael, @AustinJustice, @WHFraudTF, @ns123abc). French politics, taxpayer-fraud press videos, Elon "boss battle traffic" clips, Naval-on-Musk anecdotes, Macron quantique announcement, and a crime blotter post. Skip.