1036 points · vimarsh6739 · comments

# Summary Thinking Machines released Inkling, a multimodal open-weights model that prioritizes customizability and fine-tuning capability over benchmark performance, positioning itself as a strong American competitor to Chinese open models like DeepSeek and GLM. The community sees promise in its practical utility and business model, though some question the gap between "open-weights" and truly "open-source" and want more evidence of its fine-tuning advantages.

705 points · bilsbie · comments

# Summary The discussion broadly accepts the study's finding that sleep regularity predicts mortality but emphasizes the correlation is likely confounded by underlying lifestyle factors, mental health, genetics, and socioeconomic conditions rather than being directly causal. Many commenters share personal experiences with sleep solutions while cautioning that the study's low R² values and unmeasured variables (occupation, stress, health conditions) limit its practical utility for individual decision-making.

542 points · youngtaff · comments

The discussion covers concerns about SpaceX's bond decline and IPO, with critics arguing Elon Musk's history of broken promises justifies skepticism, while supporters contend a 10% post-IPO dip is normal and SpaceX's strong business fundamentals don't support "junk bond" characterization. The broader debate reflects frustration with IPO dynamics that allegedly favor early investors over retail buyers.

484 points · skp1995 · comments

# Summary Grok Build's open-source release is viewed as a tactical response to privacy concerns and low market share, with community appreciation for the openness tempered by skepticism about the massive codebase (~1.3M lines of Rust), disabled issues/discussions, and unverified claims about data deletion. The move has already spawned numerous forks addressing privacy, multi-provider support, and UI improvements.

335 points · apsec112 · comments

# Summary The discussion praises TurnTrout for quitting Google DeepMind over ethical concerns about AI weapons development, but many commenters express cynicism that individual principled stands are futile against fundamentally corrupt megacorporations that prioritize profits over ethics commitments. Some debate whether smart AI weapons are actually unethical and question why the author didn't engage with existing activist groups or address Google's broader unethical practices.

301 points · gnyeki · comments

# Summary The proposal to add Rust-style editions to SQLite is met with mixed reactions: supporters appreciate the backwards-compatible way to modernize defaults, while skeptics argue that SQLite's current design choices are deliberate trade-offs suited to its embedded use case, and that users should either accept these defaults, use compile-time options, or switch to a different database if they need different behavior.

121 points · htrp · comments

The discussion is minimal, with one commenter pointing out this is a duplicate post, another sharing the model card link showing it's a 975B parameter model with 41B active parameters, and a flagged comment.

200 points · evakhoury · comments

The discussion centers on differing philosophies for building AI agent harnesses: some commenters advocate for deterministic, Unix-philosophy-inspired systems with LLMs handling only edge cases, while others argue for simpler approaches or domain-specific harnesses over generic solutions. There's skepticism about the "do anything" claim and debate over whether complex harnesses actually outperform simpler tools like Claude Code.

283 points · davidbarker · comments

# Summary Users are largely skeptical of the $230 Codex Micro device, questioning its value proposition over cheaper alternatives like Stream Decks ($130) and pointing out it's just a rebranded Work Louder keyboard that's already outdated since Codex was folded into ChatGPT. The consensus is that it's an expensive specialty product solving no clear problem, viewed as either a misguided marketing stunt or a provocative statement about the future of work.

94 points · eieio · comments

# Summary The discussion critiques both the article's framing (blaming Claude rather than the user's manipulation) and a broader pattern of irresponsible AI agent deployment, with commenters emphasizing that LLMs should never be trusted with unsandboxed access to sensitive systems.

99 points · tony_cannistra · comments

# Summary The discussion is deeply divided: some commenters support the studio's human-centered approach and value craft, while others dismiss it as impractical gatekeeping, arguing AI is just a tool like any other and that the site's downtime undermines their credibility. The tension reflects broader disagreement about whether rejecting AI is principled or simply unrealistic.

91 points · ohong · comments

The comments largely dispute the article's significance, noting that 105 founders out of 13,000 YC alumni is a tiny sample that doesn't prove much, while also attributing the concentration to Sam Altman's existing connections with YC founders rather than any broader trend.

38 points · nkov47 · comments

Coasty's main strength appears to be blending deterministic workflows with AI-assisted recovery for unexpected situations, rather than pure agent-based automation—a differentiator the founders undersold in their pitch. Key technical challenges remain around verifying form state changes beyond visual feedback and seamlessly handing control between deterministic scripts and AI agents.

117 points · chaosharmonic · comments

# Summary Bluesky acquired the AT Protocol trademark to protect the protocol from legal threats, with plans to transfer it to an independent governance organization eventually. Commenters debate whether single-vendor control of standards is problematic, with some praising the move while others express concerns about centralized ownership.

15 points · tosh · comments

No comments available.

7 points · mfiguiere · comments

No comments available.

258 points · theanonymousone · comments

118 points · r2sk5t · comments

258 points · levmiseri · comments

70 points · dsego · comments

115 points · kadhirvelm · comments

51 points · ibobev · comments

64 points · janandonly · comments

90 points · pilililo2 · comments

47 points · omgmog · comments

41 points · felixdoerp · comments

62 points · ninglor · comments

134 points · spacemarine1 · comments

169 points · treve · comments

214 points · coolelectronics · comments

44 points · BirAdam · comments

43 points · Cider9986 · comments

53 points · johnbarron · comments

116 points · SirOibaf · comments

177 points · dxs · comments

194 points · soheilpro · comments

218 points · bilsbie · comments

60 points · andsoitis · comments

105 points · serial_dev · comments

18 points · derbOac · comments

163 points · ristello · comments

37 points · terrablue · comments

77 points · hmate9 · comments

17 points · downbad_ · comments

37 points · fabioricardo7 · comments

16 points · linolevan · comments

124 points · benswerd · comments

245 points · hermanzegerman · comments

36 points · abduhl · comments

94 points · lsferreira42 · comments