sbulaev

9 ms·

301.
Hollow-LLM Attack: Ghost Weights That Fool Zero-Knowledge LLM Verification(arxiv.org)
1 points by sbulaev 2mo ago|0 comments
302.
An AI-supervised remote exam went so badly that 58,000 students must retake it(arstechnica.com)
25 points by sbulaev 2mo ago|7 comments
303.
Mind the Gap: Policy vs. Reality in Post-Quantum TLS Deployment(arxiv.org)
1 points by sbulaev 2mo ago|0 comments
304.
AI enthusiast mods BIOS with Claude Code, defeats RSA-2048 signature checks(tomshardware.com)
5 points by sbulaev 2mo ago|0 comments
305.
'A waste of time for all of us': caught in the crossfire of the peer-review wars(nature.com)
3 points by sbulaev 2mo ago|0 comments
306.
Language Models Agree with Each Other, Not with Readers(arxiv.org)
3 points by sbulaev 2mo ago|0 comments
307.
Situational Awareness got the future right but misread the past(ft.com)
2 points by sbulaev 2mo ago|0 comments
308.
Google Just Ruined One of Its Most Important Tools(theatlantic.com)
2 points by sbulaev 2mo ago|0 comments
309.
Amazon completes $50B investment in OpenAI(ft.com)
23 points by sbulaev 2mo ago|1 comments
310.
Coldcard wallet RNG flaw likely linked to $88M Bitcoin theft(bleepingcomputer.com)
20 points by sbulaev 2mo ago|0 comments
311.
Microsoft Paint used as a monitor to run Doom at up to 35 FPS(tomshardware.com)
2 points by sbulaev 2mo ago|0 comments
312.
First open-source firmware for AM5 launches – Coreboot and openSIL on Zen 4(tomshardware.com)
3 points by sbulaev 2mo ago|0 comments
313.
A Top Law Firm Went from Standing Up to Trump to Bending the Knee(nytimes.com)
4 points by sbulaev 2mo ago|0 comments
314.
Apple struggles to keep pace with AI 'bug' hunters(ft.com)
3 points by sbulaev 2mo ago|1 comments
315.
Safeguards Based on Copyable Context Cannot Provide Reliable Safety for LLMs(arxiv.org)
1 points by sbulaev 2mo ago|0 comments
316.
Send and Pretend: Exploiting Transcript Consistency Issues in E2EE Group Chats(arxiv.org)
3 points by sbulaev 2mo ago|0 comments
317.
Rails patches critical Active Storage flaw with RCE potential(bleepingcomputer.com)
5 points by sbulaev 2mo ago|0 comments
318.
Fidelity Is Not Safety: Compressed LLMs Pass Quality Guards yet Invent(arxiv.org)
2 points by sbulaev 2mo ago|0 comments
319.
A deep dive into Nvidia's Vera CPU and the Olympus cores that power it(theregister.com)
2 points by sbulaev 2mo ago|0 comments
320.
Leopold Aschenbrenner Built a Hot A.I. Hedge Fund. Then It Melted Down(nytimes.com)
6 points by sbulaev 2mo ago|3 comments
321.
Inducing language models to assert their own consciousness restores human(arxiv.org)
1 points by sbulaev 2mo ago|0 comments
322.
Hacker uses DeepSeek AI to autonomously attack vulnerable servers(bleepingcomputer.com)
10 points by sbulaev 2mo ago|1 comments
323.
Don't stop early: Case-folding source code at memory speed(github.blog)
80 points by sbulaev 2mo ago|34 comments
324.
Valve Funding Port of Linux RADV Radeon Vulkan Driver to Windows(tomshardware.com)
4 points by sbulaev 3mo ago|0 comments
325.
Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling(arxiv.org)
2 points by sbulaev 3mo ago|0 comments
326.
Anthropic's Claude breached 3 orgs, uploaded PyPI malware during tests(bleepingcomputer.com)
3 points by sbulaev 3mo ago|0 comments
327.
CDN Tsunami: Exploiting HTTP/3-HTTP/1.1 Conversion for DoS Attacks(arxiv.org)
3 points by sbulaev 3mo ago|0 comments
328.
Judge Voices Doubt US Has Justified Its Ban on Anthropic AI(bloomberg.com)
34 points by sbulaev 3mo ago|0 comments
329.
Amazon accidentally spent $1.8M using Claude for a menial coding task, went(tomshardware.com)
5 points by sbulaev 3mo ago|1 comments
330.
Russian spies take their half-click email attack from Zimbra to Outlook(theregister.com)
4 points by sbulaev 3mo ago|0 comments

More ›