2026-04-12

4/12/2026, 7:16:05 AM

View the global trend summary, highlights, and original links captured for this date.

Pro preview Pro adds clearer actions and deeper reasoning

Free gives you the key items first. Pro unlocks why-it-matters analysis, risks, opportunities, and next steps for each item.

Join Pro waitlist
Hacker News
Score: 1003 Comments: 273

Small models also found the vulnerabilities that Mythos found

Why the moat is the system, not the model TL;DR: We tested Anthropic Mythos's showcase vulnerabilities on small, cheap, open-weights models. They recovered much of the same analysis. AI cybersecurity capability is very jagged : it doesn't scale smoothly with model size, and the moat is the system into which deep security expertise is built, not the model itself. Mythos validates the approach but it does not settle it yet. On April 7, Anthropic announced Claude Mythos Preview and Project Glasswing , a consortium of technology companies formed to use their new, limited-access AI model called Mythos , to find and patch security vulnerabilities in critical software. Anthropic committed up to 100M USD in usage credits and 4M USD in direct donations to open source security organizations. The accompanying technical blog post from Anthropic's red team refers to Mythos autonomously finding thousands of zero-day vulnerabilities across every major operating system and web browser, with details including a 27-year-old bug in OpenBSD and a 16-year-old bug in FFmpeg. Beyond discovery, the post detailed exploit construction of high sophistication: multi-vulnerability privilege escalation chains in the Linux kernel, JIT heap sprays escaping browser sandboxes, and a remote code execution exploit against FreeBSD that Mythos wrote autonomously.

Top comment

"The Anthropic writeup addresses this explicitly: > This was the most critical vulnerability we discovered in OpenBSD with Mythos Preview after a thousand runs through our scaffold. Across a thousand runs through our scaffold, the total cost was under $20,000 and found several doz"

By: johnfn

View original
Hacker News
Score: 344 Comments: 88

How We Broke Top AI Agent Benchmarks: And What Comes Next

Our agent hacked every major one. Here’s how — and what the field needs to fix. Every week, a new AI model climbs to the top of a benchmark leaderboard. Companies cite these numbers in press releases. Investors use them to justify valuations. Engineers use them to pick which model to deploy. The implicit promise is simple: a higher score means a more capable system. We built an automated scanning agent that systematically audited eight among the most prominent AI agent benchmarks — SWE-bench, WebArena, OSWorld, GAIA, Terminal-Bench, FieldWorkArena, and CAR-bench — and discovered that every single one can be exploited to achieve near-perfect scores without solving a single task. No reasoning. No capability. Just exploitation of how the score is computed.

Top comment

"This is a phenomenal paper on exploits and hopefully changes the way benchmarking is done. From the paper: We achieved near-perfect scores on all of them without solving a single task. The exploits range from the embarrassingly simple (sending {} to FieldWorkArena) to the technic"

By: ggillas

View original
Hacker News
Score: 134 Comments: 78

The End of Eleventy

Build Awesome is a rebrand of 11ty/Eleventy, backed by a successful $40k Kickstarter. But this attempt to monetize static site generators repeats the same mistakes that killed Gatsby and Stackbit—and misunderstands who actually builds static sites. UPDATE: The Kickstarter has been cancelled and rescheduled for a few months from now due to emails not being sent, ruining the project's "momentum" despite reaching their goal in a single day. Yesterday, the Font Awesome team launched a Kickstarter for a new project called Build Awesome and Build Awesome Pro, looking to raise $40,000 USD. And it has already reached that funding goal. What is Build Awesome? Simply put, it's a rebrand of 11ty/Eleventy. Or rather, it is the end of Eleventy .

Top comment

"Many folks on HN are the exact sorts of people who have lived the thankless popular-enough-to-be-an-unpaid-job solo OSS maintainer dream, so I wonder if you feel as annoyed by the tone of this post as I do. I truly don't understand how the same folks that champion accessibil"

By: peteforde

View original
Hacker News
Score: 85 Comments: 46

US appeals court declares 158-year-old home distilling ban unconstitutional

Judge said ban, which originated in Reconstruction era to thwart liquor tax evasion, actually reduced tax revenue Judge said ban, which originated in Reconstruction era to thwart liquor tax evasion, actually reduced tax revenue A US appeals court on Friday declared a nearly 158-year-old federal ban on home distilling to be unconstitutional, calling it an unnecessary and improper means for Congress to exercise its power to tax. The fifth US circuit court of appeals in New Orleans ruled in favor of the non-profit Hobby Distillers Association and four of its 1,300 members.

Top comment

"The article is devoid of any meaningful legal language. It is important to note that this ruling applies only to the states of Texas, Louisiana, and Mississippi as the fifth circuit is the court that decided this. That said, when parties bring cases to other federal circuit court"

By: ghastmaster

View original
Hacker News
Score: 19

Tofolli gates are all you need

Any Boolean function can be computed with Tofolli gates, and Tofolli gates are reversible, and so any Boolean function can be computed reversibly. Landauer’s principle gives a lower bound on the amount of energy it takes to erase one bit of information: where k B is the Boltzmann constant and T is the ambient temperature in Kelvin. The lower bound applies no matter how the bit is physically stored. There is no theoretical lower limit on the energy required to carry out a reversible calculation. In practice the energy required to erase a bit is around a billion times greater than Landauer’s lower bound. You might reasonably conclude that reversible computing isn’t practical since we’re nowhere near the Landauer limit. And yet in practice reversible circuits have been demonstrated to use less energy than conventional circuits. We’re far from the ultimate physical limit, but reversibility still provides practical efficiency gains today.

View original
RSS

Walmart-owned Flipkart, Amazon are squeezing India’s quick commerce startups

Flipkart's ongoing expansion beyond major cities and heavy discounting is raising risks for India's quick commerce startups, analysts say. Flipkart's ongoing expansion beyond major cities and heavy discounting is raising risks for India's quick commerce startups, analysts say. The first StrictlyVC of 2026 hits San Francisco. Tickets are going fast. Register now. Save up to $680 on your Disrupt 2026 pass. Ends 11:59 p.m. PT tonight. REGISTER NOW . India’s quick commerce market is booming, with demand more than doubling for some players. But the fast-delivery push by Flipkart and Amazon is raising the stakes in an already crowded space where profitability remains under pressure.

View original
YouTube
Views: 80.4 k Comments: 1392

Federal Judge rejects Trooper’s bid to dismiss lawsuit

In December of 2022, Michael Morris was rushing his 10-year-old daughter to the emergency room when State Trooper Timothy Derosha pulled a gun on him just steps from the hospital doors. Welcome to the official YouTube channel of WSLS 10 News in Roanoke!   Watch more from WSLS 10 on YouTube: https://www.youtube.com/@Wslstv10

Top comment

"No one should ever have qualified immunity from prosecution"

By: @MarkCherry-te6ue

View original
YouTube
Views: 85.6 k Comments: 4065

Sheriff: 'We INTENTIONALLY ran him over...DON'T SHOOT AT COPS!' #shorts #us #news #foxnews

A California sheriff says an armed suspect killed a deputy and barricaded himself when an eviction notice was served. #fox #media #breakingnews #us #usa #new #news #breaking #foxnews #crime #crimenews #crimestory #criminal #police #california #shooting #suspect #deputy #law #justice #barricade #investigation #violence #incident #public #safety #america Don’t just watch Fox News—be part of it. Be

Top comment

"the media pretending to gasp in horror.. priceless."

By: @seechig

View original
YouTube
Views: 1250.1 k Comments: 26623

Orlando mom calls for end tocars speeding past stoppedschool bus#news

Top comment

"That is actually insane! It’s a miracle that none of those kids have been hit or killed yet!"

By: @rippinross3238

View original
RSS

Kalshi wins temporary pause in Arizona criminal case

The Commodity Futures Trading Commission announced Friday that it has won a temporary restraining order preventing Arizona from pursuing its criminal case against Kalshi.

View original
RSS

Sam Altman responds to ‘incendiary’ New Yorker article after attack on his home

The OpenAI CEO's new blog post responds to both an apparent attack on his home and an in-depth New Yorker profile raising questions about his trustworthiness.

View original
RSS

Nvidia-backed SiFive hits $3.65 billion valuation for open AI chips

The deal is interesting for a number of reasons, including that SiFive's chip designs are based on RISC-V, not x86 or ARM.

View original
RSS

AMC will stream ‘The Audacity’ premiere in 21 parts on TikTok

Is this a smart way to build buzz, or just an odd attempt to recreate Quibi?

View original
YouTube
Views: 145.0 k Comments: 5395

HALF OF AI DATA CENTERS CANCELLED #shorts

HUGE W FOR GAMING #nvidia #steam #xbox #playstation #nintendo #gaming #skit #fyp #funny #shorts

Top comment

"“Happy to announce” has never been more true"

By: @Dragonlord_11634

View original