1034 points · aziaziazi · comments

# Summary While commenters welcome France's ban on unsolicited telemarketing calls, most are skeptical about its effectiveness without proper enforcement and technical solutions like caller authentication and carrier-level filtering—noting that existing laws in other countries haven't stopped spam calls from scammers using spoofed numbers and illegal intermediaries.

640 points · quantumgarbage · comments

# Summary Researchers demonstrated they can extract proprietary reasoning traces from frontier LLM APIs by replaying them into weaker, more easily jailbreakable models—a vulnerability the providers have since patched. The discussion centers on whether this constitutes "stealing" (many argue you can't steal what you paid for or outputs from public training data) and whether encryption and obfuscation are adequate security measures for protecting model reasoning processes.

571 points · nikolay · comments

# Summary The discussion explores the thesis that compression and prediction are fundamentally equivalent, grounded in information theory, while debating nuances around generalization, the sufficiency of this framework, and its applications to LLMs and other statistical models.

545 points · stevekemp · comments

# Summary The discussion celebrates England's progress toward eliminating hepatitis C through improved screening and treatment, with commenters noting the stark contrast to the NHS's past infected blood scandal. Several jokes reference Rust programming language rewrites, while others raise questions about the article's clarity regarding specific elimination targets.

459 points · ilamont · comments

# Summary The discussion reflects widespread skepticism that ethics departments at AI companies are structurally ineffective—viewed as PR exercises that lack real influence and are inevitably marginalized when they conflict with business priorities. Multiple commenters argue that ethics cannot be outsourced to a single team; without organizational commitment and consequences for bad decisions, ethics roles become obstacles to be worked around until they're quietly disbanded.

403 points · RickJWagner · comments

# Summary The discussion centers on shrinkflation—the widespread practice of reducing product sizes while maintaining or increasing prices—with commenters sharing examples from candy bars to household items. While some argue shrinkflation is inevitable given inflation and commodity costs, and potentially beneficial for reducing junk food consumption, most view it as deceptive corporate practice that cheats consumers.

396 points · 0xedb · comments

# Summary The discussion reveals sharp disagreement about Go's suitability for LLM-assisted coding. While some praise Go's simplicity, tooling, and consistency for AI generation, critics argue that languages with stronger type systems (like Rust) and better compile-time error detection provide superior guardrails for LLMs, and that Go's verbosity and weak safety guarantees make it problematic despite readability. Many commenters dismiss the article as biased Google marketing while advocating for alternative languages better suited to LLM constraints.

59 points · colesantiago · comments

The discussion appears to have been moved to another thread, with most comments flagged or removed. There's minimal substantive content to summarize.

306 points · rvz · comments

# Summary Users are impressed by Grok Bot's agent architecture and autonomous capabilities, but express serious concerns about security risks, data privacy, high token costs, and the precedent of giving corporations deep access to personal accounts and data.

295 points · frabonacci · comments

The discussion clarifies that the 11-16× speedup only applies to llama.cpp running in Apple's Virtualization.framework VMs (where a kernel selection bug was fixed), not to general Apple Silicon performance, with commenters also noting confusion about the title, curiosity about why the VM reports limited GPU capabilities, and appreciation for the broader ecosystem being built around Apple Silicon for AI despite Apple's limited support.

103 points · asdajksbda · comments

The discussion shows minimal engagement with the paper itself, instead focusing on meta-comments about idea originality, skepticism toward the claims, pricing questions, and criticism of the project name.

166 points · thm · comments

Users express skepticism about Manus's collapse, noting its brief hype cycle, mediocre product compared to alternatives, and questionable $2B valuation—with one commenter also critiquing its unfortunate branding.

24 points · advaith08 · comments

# Summary The discussion raises skepticism about AI-driven materials discovery's practical impact, highlighting that feasibility and cost matter more than theoretical plausibility, while acknowledging the core challenge is closing the computational-to-experimental validation loop that larger companies may already handle in-house.

46 points · zackashen · comments

# Summary The launch received mixed reactions: while users praised the concept and UI details, they raised significant concerns about onboarding friction (phone number requirement, sign-up bugs), video generation costs/limitations, and questioned the business viability against free LLM alternatives. There's also a potential trademark conflict with an existing keet.io domain.

12 points · iamskeole · comments

# Summary OpenAI's COO has resigned, and the company's Head of Ethics also recently departed within a year of joining, suggesting internal instability or disagreement over the organization's direction.

275 points · theanonymousone · comments

254 points · cdrnsf · comments

228 points · gregsadetsky · comments

209 points · doubletwoyou · comments

245 points · dcminter · comments

260 points · whosgotch · comments

159 points · robtherobber · comments

182 points · newsomix9xl · comments

98 points · simonebrunozzi · comments

97 points · robtherobber · comments

99 points · whosgotch · comments

79 points · skadamat · comments

59 points · rbanffy · comments

110 points · andsoitis · comments

73 points · tapper · comments

57 points · prathamtharwani · comments

77 points · ck2 · comments

48 points · master_crab · comments

47 points · cl42 · comments

43 points · CGMthrowaway · comments

61 points · cwwc · comments

151 points · michaefe · comments

72 points · Gander5739 · comments

41 points · pseudolus · comments

50 points · stevekrouse · comments

38 points · matada_ · comments

36 points · cdrnsf · comments

126 points · shagie · comments

40 points · dangle1 · comments

56 points · daau · comments

25 points · GranularRecipe · comments

58 points · Taikhoom10 · comments

24 points · alvis · comments

30 points · pluc · comments

237 points · droidjj · comments