janalsncm

4 ms·

1.
by janalsncm 5d ago
Right, if a model says it is Qwen there is no way to distinguish a ModernBert fine tuned with Qwen completion data from a Qwen model fine tuned with completion data. It’s also entirely possible that they used completions from a pool of open
2.
by janalsncm 5d ago
> employers expect health care plan costs to rise by an average of 11 percent per worker in 2027, unless benefits are cut I am sure most workers would prefer to get that extra 11% as a cash raise, but because healthcare costs are out of
3.
by janalsncm 5d ago
Imagine meeting a person who could do all of those things. “I once met a person who could beat any grandmaster in chess, translate any language, and complete international math Olympiad problems. He couldn’t solve any Millenium problems tho
4.
by janalsncm 5d ago
On your particular point about finding the most “effective” solution, this is something that I expect agents to be very good at. When AI does it we call it “reward hacking” but when humans do it we call them clever.
5.
by janalsncm 8d ago
To be clear, the claim is false as it is written as far as I can tell. That said, young people have a lot more expenses. They need to buy houses and raise kids and get bigger cars. Retirees largely don’t have to do those things. So I can im
6.
by janalsncm 8d ago
An LLM is a Swiss Army knife. This is a corkscrew. All of the other tasks a general-purpose LLM can do (write me a poem about pizza, rewrite this code in rust, tell me about the causes of the war of the roses) are unsupported. The only use
7.
by janalsncm 8d ago
Hey, I’m really happy that someone is building this. I tried doing something similar a couple of months ago and came to the conclusion that the dataset was at least as important as the modeling itself. Building a good dataset is nowhere nea
8.
by janalsncm 8d ago
> average pension in France is now higher than the average salary This sounded crazy so I looked it up. I don’t think it’s quite true, but after you subtract taxes and living expenses (a lot of old people own their homes) and add investm
9.
by janalsncm 8d ago
The typical pattern is called “deskilling”. It doesn’t usually mean a skilled profession will disappear overnight. Instead, the job might be done by less expensive folks like paralegals. An example is in the banking industry, where making a
10.
by janalsncm 8d ago
The point is that OS specific licenses shouldn’t exist. Especially when Microsoft makes Word and Windows.
11.
by janalsncm 9d ago
PTQ and vector quantization aren’t used for this because part of the point of ternary LLMs is to make them faster. In a ternary LLM every weight is an add, subtract, or no-op so it is fast on CPU. If you’re just using a code book to reconst
12.
by janalsncm 9d ago
Ok, but it’s a coffee shop. An Americano looks the same as drip coffee. A cold brew looks the same as an iced americano.
13.
by janalsncm 9d ago
I agree that it wasn’t a strong response, but 1) we can still be nice and 2) if you read the rest of their comment they do have substantive things to say.
14.
by janalsncm 9d ago
> I think there's still quite a huge gap between optimized and perfected version of thing that exists versus thing that entirely does not exist Neither the images created by ChatGPT nor an artists rendition of a latte exist, but bas
15.
by janalsncm 9d ago
The relevant question is how much comments on Instagram actually translate to changes in sales. If I’m going to a coffee shop, I might not like the AI coffee sign, but it’s much more important to me how good the coffee is and how expensive
16.
by janalsncm 10d ago
Technically “new user” I think.
17.
by janalsncm 10d ago
To directly answer your question, my guess is they have an alert set up for various places on the internet, and are able to post the links faster than anyone else. Maybe that part is automated. AI product updates probably tend to rank very
18.
by janalsncm 10d ago
I would definitely take you up on that.
19.
by janalsncm 10d ago
If something has general intelligence it should be able to read the rules of a game and follow them. Therefore an artificial general intelligence (AGI) should be able to do this. So we have a situation where very powerful and influential pe
20.
by janalsncm 10d ago
> It's nothing like a traditional LLM and so should not be compared to one. Frontier LLMs are expensive jack of all trades. You can absolutely compare them to purpose-built tools on any domain they touch. Engineering is all about as
21.
by janalsncm 10d ago
Idk, these seem pretty innovative: https://www.dezeen.com/2026/02/17/chinese-projects-completed... Not that I like all of them, but the point stands.
22.
by janalsncm 10d ago
That doesn’t explain why the architecture students in the article prefer facades that the general public generally dislike.
23.
by janalsncm 10d ago
Large language models are not the only type of model.
24.
by janalsncm 10d ago
Technically speaking when you send the prefix “The capital of France is “ into an LLM it will also produce probabilities across its whole vocabulary.
25.
by janalsncm 10d ago
I don’t think it’s misleading if you compare on the use cases they suggested. It’s faster and cheaper (no idea if higher quality), so it’s immediately interesting for certain things. And if you buy their RLCD claims, this might be even bett
26.
by janalsncm 13d ago
So glad you shared this talk. Having people like Bruce Schneier around in a time like this is really a gift. For those who haven’t watched, his breakdown of types of “hacking” is really good.
27.
by janalsncm 13d ago
Yoshua Bengio is a brilliant researcher who contributed enormously to earlier development of artificial intelligence. But with this sentence, > They took actions that would be considered as crimes if a human took them He is so close to t
28.
by janalsncm 13d ago
How hard is it to flash the ROM of these TVs with something that doesn’t spy on people? Feels like in the age of Claude Code the answer should be to strip out the evil side of the Faustian bargain of adtech-subsidized smart devices. Hackers
29.
by janalsncm 13d ago
Jevon’s paradox is a common talking point but it is not a law of nature. LED lightbulbs use 80% less energy than incandescent but you don’t see people using 5x more lights on their homes. The overall energy used to light homes has decreased
30.
by janalsncm 16d ago
> When people believe that AI could figure this out on its own, without first running extensive and expensive basic science experiments, they have left the realm of science and have entered what can only be described as religion. I stron

More ›