jmalicki

16 ms·

301.
by jmalicki 5mo ago
I won't talk about Lion's Mane specifically, but there is a ton of science about things that induce neurotrophic growth factor, which some claim Lion's Mane to do, and research in concussions, depression, and Alzheimer's
302.
by jmalicki 5mo ago
There are two answers already and neither is entirely adequate. In normal LLM training, you take a set of documents and have it learn to predict the future, then have some private RLHF/RLVR etc. data that it learns to produce good chat
303.
by jmalicki 5mo ago
That's not lack of knowledge of a specific language, that's just not knowing how to code period. Anything in any language where it is trying to change 40 files for something simple should have you stopping.
304.
by jmalicki 5mo ago
A lot/most of the neural regeneration drugs are still very much on patent, and very very very expensive, and a lot of insurance doesn't cover them. While there may be some conspiracy stuff there, "big medicine" saying yo
305.
by jmalicki 5mo ago
That would be bad, which is why Rust is preferred to python. In Rust when things break the AI gets a clear error message that makes it clear how to fix. In Python when things break, the AI will randomly spin its wheels for days trying dif
306.
by jmalicki 5mo ago
If the idempotency key must be persisted, that is the bookkeeping that has to occur to consider the request a success.
307.
by jmalicki 5mo ago
This talks about connections. My understanding is that while brain cell death (outside of the hippocampus, at least) cannot regenerate, the connections and networks can. But neurons regenerating connections between each other is, afaik, bee
308.
by jmalicki 5mo ago
And if there is part of the process that isn't idempotent, you want to make that surface area as small as possible , so that only failures at that one discrete step can cause issues.
309.
by jmalicki 5mo ago
The requester doesn't want to know which request succeeded, because they are duplicates and one is a retry! When you are using TCP, and you send the same data twice because of a delayed ack, you likewise don't care if the ACK is f
310.
by jmalicki 5mo ago
Your database should not allow both commits to happen - one should get rolled back. If it processed 99% of the request and the final bookkeeping failed because of a duplicate, that's still a failed request. Arguably this should be the
311.
by jmalicki 5mo ago
Fsync on PLP drives isn't strictly a NOP - you still take a latency hit from the round trip of the command to the NVMe device, where it is implemented as a NOP.
312.
by jmalicki 6mo ago
Small, but exemplary, and given the context of the post and arguments, developing exemplary leading companies is a sign that Poland is developing a deep base beyond mere outsourcing, even if there are a small number and they're not hug
313.
by jmalicki 6mo ago
Minor nit, but ECG and EKG are synonyms - the K is used because it is less likely to be confused verbally or in a written order... perhaps you mean EEG?
314.
by jmalicki 6mo ago
The comments have already mentioned CD Projekt and elevenlabs which I would call world class companies even if not huge scale.
315.
by jmalicki 6mo ago
Samsung, Hyundai, or Huawei never happen without starting as FDI or cheap outsourcing.
316.
by jmalicki 6mo ago
It doesn't give enough recognition to the true longest game of space kickball.
317.
by jmalicki 6mo ago
I love seeing a Shewchuk citation other than my ML background of learning conjugate gradient! He is truly a great educator!
318.
by jmalicki 6mo ago
Context windows don't necessarily cleanly divide. Getting each agent to be able to task within a context window is a hard problem. It's like like if your context window with one agent is n, your context window with 10 agents is n
319.
by jmalicki 6mo ago
That makes it not a context window. How to organize code like you said, and how agents interact with it, to keep the actual context window small is the fundamental challenge.
320.
by jmalicki 6mo ago
It's almost as if factions are made up of different people with different opinions in a loose alliance. But nah, clearly they're all goombas.
321.
by jmalicki 6mo ago
You can say anything you want in forward looking statements in the US, within reason. "They will have $1B in revenue by year end" is perfectly fine to completely make up.
322.
by jmalicki 6mo ago
What's a country with a more effective system? Not saying that the lack of a more effective system means the US's is optimal, but the outcomes for capital reallocation are far better in the US than the UK for instance. In the US,
323.
by jmalicki 6mo ago
ML run a GPU that uses matrix multiplies isn't deterministic unless you go through great pains to lock things down at the expense of performance.
324.
by jmalicki 6mo ago
Formal logic AI systems have existed and were popular in the 1980s. One of the problems is that they don't work - in the real world there are no firm facts, everything is squishy, and when you try to build a large system you end up ma
325.
by jmalicki 6mo ago
Some AI systems have done things like hack out of a docker container to access correct answers while being benchmarked. That is mildly concerning and I will give holding the AI accountable to some degree when it is actively being malicious
326.
by jmalicki 6mo ago
Arterial blood gas. Calcium score (may not count as a lab). Skin biopsy for cancer (does it count as a lab?). I'm unaware how to order my own troponin if I think I've had a heart attack (not that that's one I should DIY dia
327.
by jmalicki 6mo ago
During COVID a lot of empty flights flew because otherwise the airline could lose the gate slots.
328.
by jmalicki 6mo ago
Even with this debug mode, a different batch size can give different results for the same input - e.g. your tensor multiplies might use different blocking, hence different associativity. I posted that to show that at a bare minimum, there i
329.
by jmalicki 6mo ago
> > you can't order many of your own labs > Really? Which ones? There are extremely short lists of labs you can order yourselves. Virtually all of them are not on those short lists?
330.
by jmalicki 6mo ago
Why is there an expectation that it should be a required feature?

More ›