nostrebored

5 ms·

31.
by nostrebored 2mo ago
The idea of distance thresholding models that are trained to satisfy an ordering constraint is a bit strange. The reason it's hard is that there isn't a threshold! You can slice and dice it a ton of different ways, but the signifi
32.
by nostrebored 2mo ago
Professors have no reputation for writing good code. Academic code is fairly notorious for being under abstracted, hard to follow, messy, etc. If a passerby told me this story I’d have the same guess though!
33.
by nostrebored 2mo ago
why?
34.
by nostrebored 2mo ago
it sounds like a bad abstraction. ime llms are really bad at reasoning through efficiency and utilization problems. they fail at scheduling and bin-packing for the same reason. it's actually difficult, multivariate, and not great. i wo
35.
by nostrebored 2mo ago
It was sarcasm. It's also not particularly insightful. "AI data centers" get immense scrutiny right now. Astoria has a similar generation capacity in a much denser (and human populated!) area.
36.
by nostrebored 2mo ago
"Small" was a poor choice of words here, "low compute budget" is more what I'm getting at. In voice interactions, ttfat is actually relatively important. If you look at models with a <1s ttfat you eliminate almos
37.
by nostrebored 2mo ago
For customer support I don't think models have gotten better since gpt-4.1. The class of small models, with limited to no reasoning, that need to handle a complex issue with a touch of empathy, has not improved much. I think most are a
38.
by nostrebored 2mo ago
but... why? you almost never want these tools in any given session, and creating any of them can be done simply with Claude. looking at the PRs, it seems like that is exactly what's happening. i guess this is my generic problem with MC
39.
by nostrebored 3mo ago
we're always in uncharted territory. that's why most HN commenters think that this is hyperbole.
40.
by nostrebored 3mo ago
It’s in the weights. Context needs to be attended to to create a response, and the weights dictate what response is decoded. If you include retrieved context that has an American perspective, I imagine the think trace has some reconciliatio
41.
by nostrebored 3mo ago
it is pretty good at instruction following and has extremely fast decode.
42.
by nostrebored 3mo ago
qwen3.6 starts at 27B
43.
by nostrebored 4mo ago
Oh I agree with you, it's just that I don't think LLMs are either. If you think of LLM knowledge, especially in scientific/engineering fields, as a lossy representation of the density of ideas, then you'd expect to see s
44.
by nostrebored 4mo ago
Daily allergy tablets are associated with huge increases in early onset Alzheimer’s. Glad you found something that works, but might be good to get some of the allergen injections :)
45.
by nostrebored 4mo ago
I don’t understand the negative reactions. Medical care as it exists requires the doctor and patient to have their brains switched on. I’ve almost never had a problem where a doctor provides me with a diagnosis and I go about my day. Most o
46.
by nostrebored 4mo ago
Claude is the worst FM at image understanding. Prior to gpt-5.4 the only usable models were Gemini and Qwen.
47.
by nostrebored 4mo ago
But the binary classification models can be made ternary easily. RL on congruence plus penalty for misdiagnosis is easy to set up and gives great results. What I’ve seen be the true bottleneck is people not setting up the structured data. B
48.
by nostrebored 4mo ago
I think that much of the visual gap is because what to attend to in images is less structured. Anecdotally small qwen finetunes (ie less than 10B) take task accuracy from sub 30% on FMs to 90%. We have sold some of these for outcome based b
49.
by nostrebored 4mo ago
He made it happen by continuously using doomsday marketing to pump up model capabilities. This is the comeuppance. There is a huge contingent of people who do not interact with AI on a daily basis. Many of them legitimately believe we'
50.
by nostrebored 4mo ago
Well if you can do this then you don't delegate execution path derivation to the agent. The benefit is a predictable coherent world state where you understand the impact of { current state } x { action } without having to enumerate tha
51.
by nostrebored 4mo ago
The leakage problem is so pervasive. None of the frontier models seem to have any idea how to actually hold out rows. God help you if you decide to change the data mix. I was working on creating a next-n-actions predictor for one of our use
52.
by nostrebored 4mo ago
Building an ontology of how people think of and organize information, processes, and actions is not solved via markdown. It’s not well solved intra company much less inter. The systems that do solve some of this are optimizing for unstructu
53.
by nostrebored 4mo ago
I strongly prefer codex. Claude is annoying. Codex provides descriptions where I want them and more touchpoints to audit the quality of work. Claude code on experimental seems to not even show diffs when asked anymore, and it's much le
54.
by nostrebored 4mo ago
But how many plugins are people actually using? I can think of one MCP server I find valuable (context7) and one plugin that i've installed, but continuously think about uninstalling (obra/superpowers). Both were trivial to set up
55.
by nostrebored 4mo ago
LLMs are obsessed with “gates”. Freezing the gates here is intuitive to me as this point — don’t let validation drift.
56.
by nostrebored 4mo ago
But why does your agent control doneness? It seems to me the most odd part to delegate. All LLMs are terrible at it. Most LLM tasks can be expressed as a DAG or DAG of DAGs. Why delegate that to a random point in context instead of enforcin
57.
by nostrebored 4mo ago
i think the reason for your deluge of downvotes is that a society the promotes more things becoming affordable is one that prioritizes stability. universal access to housing, healthcare, and education that people want is only possible in a
58.
by nostrebored 4mo ago
Do you not think that the allocation of human time is one of the world’s biggest problems?
59.
by nostrebored 4mo ago
I don’t think you have ever actually lived in one. “which plays on this general tendency of humans to prefer comfort over challenge, confirmation over rejection.” This is completely and observably wrong. It reads like it would make sense, b
60.
by nostrebored 4mo ago
this is true of most actual clouds/neoclouds. oversubscription and intelligent placement of workloads is already something they do. I’ve known a few people at AWS who have offset unbelievable amounts of cost by optimizing placement.

More ›