659 points · CoryOndrejka · comments

Mike Tomlin, the former Pittsburgh Steelers head coach, spent 12 years building an elaborate Minecraft city on a console, impressing people with his dedication, architectural taste, and the earnest joy he brings to sharing his creative project—a heartwarming story that contrasts sharply with concerns about AI-generated content and highlights how gaming transcends celebrity culture.

621 points · bastitx · comments

# Summary Kolibri, a new open-weight German-focused LLM from Aleph Alpha (soon merging with Cohere), is praised for its transparency and suitability for European business use cases like document processing, but criticized for being outperformed by Qwen3.8 27B and for questionable "sovereignty" claims given the Cohere acquisition and reliance on other LLMs in its training pipeline.

468 points · elffjs · comments

# Summary The discussion reveals that hard budget caps on cloud services are critically needed but deliberately absent from most US providers' offerings—likely by design to maximize revenue—while European providers and some newer services have implemented them successfully, suggesting the technical barriers cited are mostly excuses.

448 points · azhenley · comments

# Summary Commenters share nostalgic memories of Newgrounds as a formative creative and gaming platform from their youth, celebrating its preservation of Flash content through Ruffle emulation and its enduring role as an independent, community-driven alternative to commercialized modern platforms like Roblox and YouTube.

437 points · sbulaev · comments

# Summary Commenters debate whether Flock's license plate readers constitute unconstitutional mass surveillance, with critics warning of dystopian tracking capabilities and government abuse, while defenders argue the technology is necessary for crime prevention and operates on public roads where privacy expectations are minimal.

413 points · tejaskumar__ · comments

# Summary Germany's Kolibri AI model generated discussion about whether it qualifies as a "Show HN" post (most commenters said no), with interest in its technical features like the Merlin-Arthur protocol and questions about why it doesn't adopt DeepSeek's KV cache optimizations. An insider noted the model performs well on coding and agentic tasks despite being from a newly-formed team with plans for rapid iteration.

53 points · dblack12705 · comments

# Summary The discussion questions whether Jev (an AI model) is well-calibrated, with commenters arguing that its poor performance on underspecified theoretical questions like the Maxwell-Boltzmann distribution doesn't reflect poor calibration but rather the nature of unanswerable prompts—similar to asking humans to generate random distributions. Some point out Jev appears well-calibrated on practical NLP tasks and that its limitations may stem from being a small, fast model rather than fundamental calibration issues.

221 points · saikatsg · comments

# Summary Opus 5.5 is widely praised as a powerful model excelling at complex tasks like CI optimization, frontend design, and long-running projects, though users debate best practices around prompt specificity, token efficiency, and concerns about overeager behavior and spurious safety refusals.

263 points · jethronethro · comments

# Summary A safety leader at OpenAI has resigned and publicly criticized the company's culture as "broken," with the discussion linking to an earlier HackerNews thread on the same topic.

21 points · aray07 · comments

# Summary The discussion largely rejects the article's claim that manual coding is obsolete, with commenters arguing that writing code provides intrinsic satisfaction, builds algorithmic thinking, and ensures code comprehension—while questioning whether AI-generated code's non-determinism, costs, and maintenance challenges truly make it a wholesale replacement for human programming.

11 points · ChrisMarshallNY · comments

No comments available.

265 points · Brajeshwar · comments

The discussion reflects widespread skepticism about OpenAI's safety culture, with commenters debating whether the exodus of safety-focused employees represents genuine concerns about AI development risks or performative activism by people profiting from the company. Key tensions emerge around whether adequate safety standards can exist under "move fast and break things" incentives, and whether dissenters should be more credible by forfeiting their financial gains.

18 points · sensanaty · comments

# Summary Commenters are divided: some dismiss the allegations as unreliable given that two law firms withdrew representation and lack of corroboration from family members, while others argue Altman's documented behavior and statements warrant taking the claims seriously.

225 points · rbanffy · comments

209 points · timpera · comments

151 points · mooreds · comments

303 points · trwhite · comments

109 points · yu3zhou4 · comments

88 points · simonjgreen · comments

60 points · samch · comments

43 points · tty456 · comments

58 points · TeaVMFan · comments

31 points · measurablefunc · comments

41 points · tukunjil · comments

25 points · theanonymousone · comments

163 points · vinhnx · comments

34 points · xnx · comments

16 points · thelastgallon · comments

15 points · husky8 · comments

15 points · mooreds · comments

13 points · surprisetalk · comments

60 points · KomoD · comments

12 points · geox · comments

106 points · bundie · comments

16 points · Betelbuddy · comments

27 points · pjmlp · comments

11 points · jjjegly · comments

11 points · obscurette · comments

178 points · romac · comments

39 points · mondainx · comments

55 points · Topfi · comments

10 points · ajoy · comments

10 points · wyldfire · comments

11 points · alexeyindeev · comments

53 points · latein · comments

101 points · JumpCrisscross · comments

69 points · elffjs · comments

145 points · geoffbp · comments

211 points · kmeh · comments

43 points · Retr0id · comments