1279 points · tedsanders · comments

# Summary OpenAI claims to have solved the Navier–Stokes Millennium Prize Problem using AI, which would be a major achievement, but the announcement is heavily overshadowed by allegations that they may have trained on competitors' work and engaged in aggressive academic conduct to scoop the result.

1038 points · sonicrocketman · comments

# Summary The discussion centers on Internet Archive's fundraising campaign, with commenters expressing mixed support: many appreciate the organization's mission and donate, while others criticize its technical issues, aggressive rate-limiting, controversial copyright activism, and lack of transparency around the 2:1 matching scheme and funding sources.

685 points · rpgbr · comments

# Summary The discussion is highly skeptical of the headline's claim that LibreOffice's download records are due to users rejecting AI features. Commenters point out that downloads have been gradually increasing over time, and attribute the surge instead to factors like LibreOffice being bundled with AI tools like Codex, increased Linux adoption, and general long-term growth—making the causality claim speculative clickbait.

589 points · yurivish · comments

# Summary The discussion reveals sharp disagreement over AI risk: skeptics argue current evidence doesn't support existential threat claims and compare the concern to other known dangers like nuclear weapons and climate change, while believers contend that rapid AI capability gains and potential dual-use applications (bioweapons, autonomous hacking) create unprecedented risks that deserve serious attention despite the difficulty of proving them empirically.

569 points · utiiiD · comments

# Summary The discussion is mixed: while some praise DeepMind's AlphaGenome Atlas as a significant scientific achievement, others question whether it represents genuine advancement or just marketing hype, with critics noting that prior models like Borzoi perform similarly and that the tool's predictive accuracy remains limited—particularly concerning its inability to reliably predict mutation effects even in simple viruses. Additional concerns center on access restrictions, commercial licensing, and the gap between making predictions more accurate versus advancing fundamental biological understanding.

542 points · fittingopposite · comments

# Summary A couple built a house optimized for LAN parties with 18 gaming PCs, cat integration, and smart home features, which sparked admiration for their project execution and lifestyle, though some commenters expressed mixed feelings about the wealth display.

353 points · vertigoruntime · comments

# Summary The discussion praises ChatGPT Images 2.5's significant speed improvements and technical capabilities, but criticizes its showcased use cases as trivial, ethically problematic (fake photos, deepfakes), and prone to subtle errors that make them unsuitable for practical applications like interior design or product imagery.

269 points · Argonautlabs · comments

# Summary A developer successfully ran a 2.8T parameter MoE model on a MacBook Pro by streaming expert weights from four SSDs, achieving 1 token/second on short prompts but 6+ minutes for prefill—demonstrating technical ingenuity with practical limitations, with most commenters questioning the utility versus the engineering effort involved.

259 points · stared · comments

# Summary The discussion explores quantization trade-offs for Qwen 3.8 27B, with consensus that 4-bit quantization maintains quality while 1-bit collapses, though commenters debate the relevance of standard benchmarks, push for 3-bit evaluation (critical for 16GB GPUs), and note that KV cache quantization and context length matter more than commonly acknowledged.

163 points · jonbaer · comments

# Summary The discussion centers on whether OpenAI independently solved the Navier-Stokes problem or improperly built upon mathematicians Tristan Buckmaster and Levent Alpoge's work, with commenters broadly criticizing OpenAI's competitive conduct, potential misuse of researcher data, and prioritization of publicity over scientific integrity regardless of the technical facts.

471 points · domhudson · comments

Users are frustrated that Claude models are unnecessarily verbose and bury important information under layers of explanation, so they've created workarounds like the "i-have-adhd" skill, though many argue the real fix needs to come from Anthropic optimizing the model itself rather than relying on downstream patches.

79 points · tosh · comments

# Summary YC's Early Access Network appears to be a curated matchmaking service between AI startups and enterprise buyers, though commenters are skeptical about whether it's genuinely beneficial or primarily serves to de-risk YC's larger seed investments while creating artificial FOMO around exclusive access to early-stage companies.

5 points · Curiositry · comments

No comments available.

22 points · meco · comments

The discussion expresses skepticism about Devin's $48B valuation, with commenters questioning its market relevance compared to competitors like Cursor and Claude, though one notes it has improved. The consensus suggests the valuation seems disconnected from actual product adoption and value.

232 points · HatchedLake721 · comments

343 points · abraham · comments

371 points · hn_acker · comments

184 points · dotcoma · comments

169 points · robinhouston · comments

411 points · tosh · comments

126 points · mazokum · comments

108 points · bookofjoe · comments

118 points · jodacola · comments

67 points · Teever · comments

367 points · cat-whisperer · comments

72 points · KingOfMyRoom · comments

56 points · geox · comments

116 points · mathgenius · comments

214 points · tosh · comments

42 points · bookmtn · comments

50 points · asib · comments

31 points · billybuckwheat · comments

31 points · cwwc · comments

260 points · Cider9986 · comments

213 points · Topfi · comments

34 points · summm · comments

23 points · tosh · comments

41 points · samclemens · comments

230 points · animeshchouhan · comments

33 points · trencedamp · comments

101 points · ibobev · comments

19 points · 0bytes · comments

122 points · progval · comments

15 points · erlend_sh · comments

86 points · theanonymousone · comments

29 points · player85 · comments

542 points · sbulaev · comments

14 points · vinni2 · comments

18 points · nreece · comments

14 points · puzzlingcaptcha · comments