stared

14 ms·

271.
by stared 8mo ago
As a counterexample, here is an example of a Singaporean officer refusing to accept a bribe, as reported by Lee Kuan Yew: https://www.youtube.com/shorts/nZv_UkMh0FA
272.
BenchPress Predicts Gemini 3.1 Pro and Claude Opus 4.6's scores within ±2 points(twitter.com)
1 points by stared 8mo ago|0 comments
273.
by stared 8mo ago
I rerun it for GPT-5.2-Codex, for high and xhigh. Finally, it matches my experience, and it is actually good (as good as the best models for localization, still impressive 0% false positive rate): https://quesma.com/benchmar
274.
by stared 8mo ago
Rerun it for "high" and "xhigh" effort settings, and GPT-5.2-Codex still get 0% false positive, while getting at the level of other best models for localization of backdoors: https://quesma.com/benchmarks
275.
by stared 8mo ago
I am scared of two things. First, people being rubber stamps for AI recommendations. And yes, it is not unreasonable that in a dire situation, someone will outsource their judgment (day). Second, someone at the Pentagon connecting the red b
276.
by stared 8mo ago
In the topic, it brought me fond memories of "Nuclear War" (1989), https://archive.org/details/msdos_Nuclear_War_1989 . Back then, it was also AI firing nukes. Just back then, AI meant simple scripts.
277.
by stared 8mo ago
If there is a VC-backed for-profit company, the core part is how much value something brings. "Safety" here works for both PR and hiring (a lot of talented engineers and researchers might flock to it), and maybe soft power for leg
278.
by stared 8mo ago
In the topic of death poems, I consider "You Want It Darker" by Leonard Cohen a masterpiece. He was 83 with terminal cancer. Yet, this song captures both his wit & spirit at its height.
279.
by stared 8mo ago
I had to check if the creator is Polish, as "ciekawa" means "interesting". But apparently, just a coincidence.
280.
by stared 8mo ago
This I know - it was a tongue on check to the claim on „foundational models”.
281.
by stared 8mo ago
Either I miss something, or is it yet another marketing approach by Stephen Wolfram? The post talks about "foundation tool", yet it offers an MCP to a proprietary app (out of miriads of such). Sure, as any other tech, Mathematica
282.
The Science Playground – Science Games and Simulations(thescienceplayground.com)
1 points by stared 8mo ago|0 comments
283.
by stared 8mo ago
Also, the Herpes family (hint: most of us have HSV-1), vide "Alzheimer's risk 10 times lower with herpes medication" https://news.ycombinator.com/item?id=17540094 and much more, https://hn.algolia.
284.
I had everything – nd suddenly it stopped making sense(userfriendlybrain.substack.com)
4 points by stared 8mo ago|1 comments
285.
by stared 8mo ago
Thanks! I was under some impression that Apple ID is needed for these to work properly. By "connectivity" I meant like syncing Photos with MacBook.
286.
by stared 8mo ago
To be honest, it is also our surprise. I mean, I used GPT 5.2 Codex in Cursor for decompiling an old game and it worked (way better than Claude Code with Opus 4.5). We tested for Opus 4.6, but waiting for public API to test on GPT 5.3 Codex
287.
by stared 8mo ago
Nice! Though, $2000 is a step price. I had some fun with using an Inkplate e-ink display - bough a bare 5" for €74 (a 10" with batteries is there €219). Smaller, but also way more affordable. It connects via WiFi, and make it disp
288.
by stared 8mo ago
How you deal with connectivity, since the newest system it can use is iOS 15, while Apple ID needs iOS 16.2 or later?
289.
by stared 8mo ago
You can solve any problem with AI if you give enough hints. The question we asked is if they can solve a problem autonomously, with instructions that would be clear for a reverse engineering specialist. That say, I found these useful for ma
290.
Ask HN: Is the original iPhone SE just a brick now?
4 points by stared 8mo ago|5 comments
291.
by stared 8mo ago
No. To give it a fair test, we didn't tinker with model-specific context-engineering. Adding skills, examples, etc is very likely to improve performance. So is any interactive feedback. Our example instruction is here: https:/&#x
292.
by stared 8mo ago
Thanks for sharing! It seems to be an active space, vide a recent MCP server ( https://news.ycombinator.com/item?id=46882389 ). I you haven't tried, recommend a lot posting it as Show HN. I tried a few approaches - http
293.
by stared 8mo ago
One of the authors here. The tasks here are entry level. So we are impressed that some AI models are able to detect some patterns, while looking just at binary code. We didn't take it for granted. For example, only a few models underst
294.
What the coming era of highly bespoke software might look like(twitter.com)
3 points by stared 8mo ago|1 comments
295.
by stared 8mo ago
Well, here goes an obligatory mention of "At the Mountains of Madness" by HP Lovecraft. (BTW: my favourite story by him.)
296.
by stared 8mo ago
Awesome soft! It works surprisingly nicely with AI agents (I mean, like Cursor or Claude Code, I don't let it run autonomously!). Here on detecting malware in binaries ( https://quesma.com/blog/introducing-binaryaud
297.
Please don't install Clawdbot [video](youtube.com)
2 points by stared 8mo ago|0 comments
298.
by stared 8mo ago
Oh, my favourite format during my undergraduate time! Most books in mathematics and physics (some old and niche) were available in the "Russian library". At the same time, I haven't yet seen DjVu used in a legit way.
299.
We hid backdoors in binaries – Opus 4.6 found 49% of them(quesma.com)
2 points by stared 8mo ago|1 comments
300.
by stared 8mo ago
Repo: https://github.com/quesmaOrg/BinaryAudit All charts: https://quesma.com/benchmarks/binaryaudit/

More ›