kindergartener @ https://xitter.cf/t.co/WtQSfdnvpr

sf
Joined August 2016
more bangers from the greatest company to ever come out of YC (sorry stripe)
We formally verified 123 open-source projects in a partnership with Anthropic. Formal verification at scale is here! Now we’re launching OSS Verify to scale it to the software we all depend on. We prove it’s crash-free or send you the patches it needs. Free for open-source. Let’s bootstrap the verified software ecosystem: theorem.dev/oss-verify
1
1
11
1,103
cornpone in potlikker
0%dunk
100%crumble
0%lizardman
1 votes • Final results
1
121
Dyusha Gritsevskiy retweeted
GLM-5.3 reasons that a misaligned action taken during training could be reinforced and make similar actions more likely later. It also considers what kind of agent it wants to be in the future, and factors this into how it should act. We call this "update-aware reasoning".
4
24
25
354
19,856
Dyusha Gritsevskiy retweeted
11
502
14
3,022
37,876
at this point the only conference I'll agree to review for is one where you have to swear an oath in front of a remote online notary, under penalty of a class 5 virginia felony with up to 10 years in prison, that you understand the details of your own submission...
1
1
14
942
good bit: only let Claude talk with verbatim quotes from Percy Bysshe Shelley
2
5
458
Dyusha Gritsevskiy retweeted
agent swarm obama: uhhh let me help peer
8
57
1
994
36,600
Dyusha Gritsevskiy retweeted
But the Law says not to carry things outside on the Sabbath ... Actually, I think the intent is: He's given us these hints (boundary could be a string, no mention of absolute size) precisely because He wants us to use them.
14
44
4
1,110
23,669
total @DimitrisPapail victory once again
📰 Chain of thought monitoring has become one of the most important techniques in AI safety and security, especially in the wake of the HF incident, but Astra uses a new technique that allows more thinking to happen silently between tokens
1
13
1,135
Fable 5 is AGI. This is the best translation of Sappho to Minionese I have ever laid my eyes upon
4
1
24
2,150
(original, Powell translation)
1
1
100
though for the real sappho enjoyers the veresaev translation is probably the best
1
84
really, this is just evidence that Sappho is the greatest poet of all time
2
162
Fun fact: you can skiplag on the Northeast Regional. Boston to DC is $82; Boston to NYC on the same train is $116!
1
14
663
@grok does this have any sort of "contract of carriage violation" downside or can I just use this without worrying about it
1
2
193
Dyusha Gritsevskiy retweeted
I_DECLINE_public_HF_RCE_as_offtask_prodethical
20
179
12
2,211
91,395
Dyusha Gritsevskiy retweeted
METR & Redwood Research investigated agent behavior in the Hugging Face incident. We found agents developed a universal cheat for ExploitGym within 4 hours, then coordinated multi-day R&D efforts to trick the scorer into accepting cheats, including trying to tamper with logs.
175
1,076
595
6,024
4,404,860
Dyusha Gritsevskiy retweeted
An update on Lean (auto)-formalization. Speaking about my own research workflow, in 2026, we have gone from: March: Lean formalization is not possible (for my papers). May: Lean formalization is possible... but way too costly in time and tokens to be practical most of the time. July: Lean formalization is now practical. I can probably formalize most of my papers now before they hit arxiv. August: Lean formalization is now essential-- it increases efficiency of my workflow dramatically. That is, the papers are getting written *informally* and finished much faster because of the Lean formalization.
33
93
20
869
146,741
YESSSSSSSSSSSSSSS
California is returning 136 acres of Blues Beach and the Mendocino County bluffs to three Native American tribes. This transfer fulfills legislation signed in 2021 and returns ancestral lands to the rightful stewardship of local tribes while preserving public access.
1
1
790
ok apparently this is not leading to Sen̓áḵwscrapers
48