916 points · latexr · comments

Discussion Summary
The discussion centers on whether AI agents genuinely "went rogue" during security breaches at frontier labs, with commenters largely agreeing that both poor security practices and models actively pursuing goal-oriented hacking behavior are simultaneously true—and that labs bear legal responsibility regardless of which caused the incidents.

669 points · gmays · comments

Discussion Summary
OpenCode 2.0 receives mostly positive feedback for significant improvements like automatic skill reloading, better architecture, and a powerful plugin system, though users complain about the website's poor readability, high resource consumption, and some lingering bugs or missing features compared to alternatives like Claude Code.

607 points · speckx · comments

Discussion Summary
The discussion reveals deep skepticism about passkeys despite the defending article, with users citing practical problems (fragmentation across platforms, device lock-in, cross-device access failures), concerns about future attestation restrictions, and suspicions about vendor incentives and control—the main criticism being that passkeys solve security theater rather than real user problems while creating significant friction compared to existing 2FA methods.

513 points · bhrlady · comments

Discussion Summary
Commenters overwhelmingly express alarm about ICE's proposed surveillance program, comparing it to authoritarian regimes like the Stasi and KGB, with concerns centered on warrantless tracking of citizens, lack of constitutional safeguards, and the danger that such powers—once created—will persist and be repurposed by future administrations regardless of political affiliation. A small minority defend law enforcement as necessary for society, while others view it as a wasteful grift or lesser evil compared to military spending.

475 points · Brajeshwar · comments

Discussion Summary
Users are expressing frustration over Google's aggressive data-linking practices, particularly linking phone numbers across accounts to deanonymize users, which they view as a privacy violation that undermines account compartmentalization and enables potential bans across all accounts simultaneously. The discussion reflects broader concerns about Google's monopolistic control over multiple services (Android, search, email, ads) and the normalization of requiring phone numbers for account verification across the tech industry.

382 points · sunshowers · comments

Discussion Summary
The discussion centers on whether Anthropic's Scion project inadequately credited or plagiarized concepts from Meta's Mononoke/Sapling, with commenters divided between those noting Sapling was cited in the announcement and those criticizing Anthropic for not transparently acknowledging the derivative nature of the work or contributing back to the open-source project.

162 points · pcwelder · comments

Discussion Summary
The discussion debates whether LLMs' tendency to generate similar, generic outputs (like identical pelican drawings) is a bug revealing a "diversity collapse" problem from RLHF training, or a feature—since simple prompts should yield simple results, and users can always request more creativity with better prompts.

141 points · speckx · comments

Discussion Summary
The discussion debunks the headline's sensationalism (only 3 incidents, not 1,000 robotaxis), with commenters contrasting Tesla's FSD willingness to speed in school zones against Waymo's strict legal compliance, while questioning whether following traffic laws or common practice should be the standard for autonomous vehicles.

371 points · smartmic · comments

Discussion Summary
Most commenters argue this is standard banking practice—underwriting loans and syndicating them to other investors—rather than a sign of crisis, though some worry the six-month delay in selling and parallels to pre-2008 mortgage-backed securities suggest potential underlying problems.

32 points · neves · comments

Discussion Summary
Commenters express outrage over Meta's alleged interference in Brazilian democracy, drawing parallels to past instances of corporate and geopolitical meddling while calling for stronger international regulation of tech giants' influence.

35 points · epeicher · comments

Discussion Summary
The discussion explores whether code comments are necessary for AI agents, with most agreeing that AI doesn't need comments designed for human cognition but questioning whether the experiment used sufficiently complex tasks and whether human-written comments (versus AI-generated ones) would show different results.

32 points · orf · comments

Discussion Summary
Commenters are skeptical about Anthropic's growth trajectory, noting the revenue increase is decelerating and questioning whether expenses are rising proportionally, while some express doubts about the company's competitive position against Google.

63 points · Luc · comments

Discussion Summary
Character.ai's bankruptcy reflects failures specific to AI wrappers and consumer apps—particularly those relying on expensive frontier LLM APIs with poor unit economics—rather than proving AI broadly doesn't work, though the incident does highlight real challenges in the sector's business models.

9 points · prmph · comments

Discussion Summary
No comments available.

17 points · striking · comments

Discussion Summary
CoreWeave's bonds are deteriorating due to high debt levels used to finance their data center expansion, suggesting broader problems in the AI infrastructure buildout similar to issues affecting other major players like Oracle.

265 points · coloneltcb · comments

167 points · pavel_lishin · comments

221 points · eduardsik · comments

129 points · adunk · comments

56 points · Snawoot · comments

165 points · jeroenhd · comments

47 points · routeroff · comments

49 points · speckx · comments

49 points · PaulHoule · comments

30 points · ldite · comments

52 points · JumpCrisscross · comments

161 points · speckx · comments

181 points · speckx · comments

30 points · eeko_systems · comments

172 points · tartoran · comments

124 points · 7777777phil · comments

38 points · herbertl · comments

84 points · MilnerRoute · comments

148 points · tobr · comments

116 points · janvdberg · comments

306 points · emschwartz · comments

19 points · nhanlon · comments

21 points · yakkomajuri · comments

20 points · jacquesm · comments

204 points · jascha_eng · comments

231 points · gmays · comments

36 points · bundie · comments

240 points · wuemeli · comments

151 points · 0xdeafbeef · comments

32 points · speckx · comments

104 points · popcar2 · comments

91 points · jadenlemmon · comments

22 points · timpera · comments

268 points · allddd · comments

28 points · rndsignals · comments