minimaltom

5 ms·

61.
by minimaltom 5mo ago
This is a great writeup! thank you!! Reminds when i did noogler training back in the day and one of the talks described a cascading failure at a datacenter, starting with a cat which was too curious near a power conditioner, and briefly con
62.
by minimaltom 5mo ago
They absolutely have backups, I presume they were ineffective or also down for _reasons_.
63.
by minimaltom 5mo ago
Worth adding also that you can only use these vectors to corrupt the page cache for files reachable in your mount namespace. Usually with containers, almost nothing is shared with the host namespaces (tho likely shared with other container
64.
by minimaltom 5mo ago
Nah, even insane token costs don't come close to the costs of labor. Most likely this is just 'AI-washing' - dressing a layoff for economic reasons (such as propping up their shrinking margins) as something more palatable to
65.
by minimaltom 5mo ago
Between this, the emotions paper, golden gate claude etc, it doesn't seem like such a stretch that Anthropic are doing some kind of activation steering as part of training (and its part of their lead)
66.
by minimaltom 5mo ago
Theres been multiple papers discussing how only updating parameters that have high agreement in update direction leads to less overfitting and better generalization. Lemme see if I can find em. https://arxiv.org/abs/241
67.
by minimaltom 5mo ago
This attack class lets you escalate from any user to UID 0. Not running as root won't save you, in fact, this attack is for those processes not running as root. However, if you are in a user namespace where UID 0 doesn't map to sy
68.
by minimaltom 5mo ago
> That is, if the batch signal on a parameter exceeds its leave-one-out noise, update it; if not, skip it. This is a one-line change to Adam that accelerates grokking by 5x, suppresses memorization in PINNs, and improves DPO fine-tuning,
69.
by minimaltom 5mo ago
my guess is they want to do AI/O as part of their event loop explicitly, and blocking a thread in a syscall waiting for an IOP (ala std::fs) isn't the vibe.
70.
by minimaltom 6mo ago
high surface energy? mahm/father I yearn for the spicy surface
71.
by minimaltom 6mo ago
Unless "with experience in the field" == academia, disagree. In particular I remember the early discourse & hype around Wireguard, it was discussed as if perfection was an achieved outcome.
72.
by minimaltom 6mo ago
The archive-handling code was in lean-zip, it just seems the verifiers forgot to write proofs for it (still a bug). Thats not the main finding of the article however. The main bug found was actually in the lean runtime, affecting all proofs
73.
by minimaltom 6mo ago
Is it fair when it comes to formally-verified software to only consider bugs that violate a proof, and ignore everything else? Formally-verified software is usually advertised "look ma no bugs!" Not "look ma no bugs*" *A
74.
by minimaltom 6mo ago
Looks like a normal distribution about the chaos mean to me. I appreciate its not everyones cup of tea, but I like this style of writing.
75.
by minimaltom 7mo ago
This 'fingerprint' function is super interesting, I imagine this is a signal they use to detect non-claude-code use of claude-code tokens: src/utils/fingerprint.ts#L40-L63
76.
by minimaltom 7mo ago
Strong agree on it all being a legal minefield / new grass. > But was almost certainly part of its training data, which complicates things On this point specifically, my read of the Anthropic lawsuit was one of the precedents was th
77.
by minimaltom 7mo ago
Is your (1) description of clean room implementation, and (2) description of what was done, actually correct? (1): my understanding was that a party _with access to copyrighted material_ made the functional spec, which was communicated to a
78.
by minimaltom 7mo ago
I feel exactly this. One thing I will add: while AI is getting really good at _doing_ the software building bits, I haven't yet seen it well integrated into the decision-making and political structure of organizations. Right now, it se
79.
by minimaltom 9mo ago
On (1): Might be useful to separate investment flows from the rest of US's economic activity. AI investment is propping up capital flows, the GDP statistic, and responsible for most of the gains on SPX, but its still a small fraction
80.
by minimaltom 9mo ago
> Does Waymo have the same object permanence and trajectory prediction (combined) to that of a human? On this note specifically ive actually been impressed, ie when driving down Oak st in SF (fast road, tightly parked cars) I've oft
81.
by minimaltom 9mo ago
There are settings to make it reproducible but they incur a non-negligible drop in performance. Unsurprising given they amount to explicit synchronization to make the order of operations deterministic.
82.
by minimaltom 9mo ago
Its not deterministic. Any individual floating point mul/add is deterministic, but in a GPU these are all happening in parallel and the accumulation is in the order they happen to complete. When you add A then B then C, you get a diffe
83.
by minimaltom 9mo ago
Some of what OP is saying generalizes to the concept of being "too early" - if you are early, your engineering / innovation spend is used to discover that at-the-time reasonable ideas don't work, or don't work with
84.
by minimaltom 9mo ago
Fair note! Just highlighting that this niche is uniquely screwed and I wouldnt wish ironing it out under the knife on anyone lol
85.
by minimaltom 9mo ago
> Anthropic was a more X-risk concerned fork of OpenAI. What is XRisk? I would have inductively thought adult but that doesn't sound right.
86.
by minimaltom 9mo ago
This is a really good question. What convinces me is this: I live in SF and have friends at various top labs, and even ignoring architecture improvements the common theme is this: any time researchers have spent time to improve understandin
87.
by minimaltom 9mo ago
Omg I love them!
88.
by minimaltom 9mo ago
There are attestation mechanisms, but huge portions of a public user-base (especially android) don't pass that check because their device is too old, or their OEM sucked, or something something mediatek SOC, or <insert esoteric deta
89.
by minimaltom 9mo ago
Fascinating! So each GPU is partnered with disk and NICs such that theres no oversubscription for bandwidth within its 'slice'? (idk what the word is) And each of these 8 slices wire up to NVLink back to the host? Feels like there
90.
by minimaltom 9mo ago
Whats a more realistic config?

More ›