potatoman22

5 ms·

31.
by potatoman22 10mo ago
At least I'm adding to the discussion.
32.
by potatoman22 10mo ago
I'm saying LLMs are deterministic and because of that, prompt engineering can be effective. You knew what I was trying to say, but chose to ignore it. You should follow the HN Guidelines. I'm trying to have a discussion, not a sna
33.
by potatoman22 10mo ago
That doesn't quite make sense to me. A meter is a unit of measure for distance, but a meter is a distance.
34.
by potatoman22 10mo ago
It's technically deterministic, but it feels nondeterministic in chatbots since tokens are randomly sampled (temp > 0) and input is varied. Using the right prompt makes the model perform better on average, so it's not completel
35.
by potatoman22 10mo ago
To be fair, discerning heteronyms might just be a gap in its training.
36.
by potatoman22 10mo ago
Ah I meant heteronyms. Thanks!
37.
by potatoman22 10mo ago
I was mistaken about what homophone means!
38.
by potatoman22 10mo ago
From what I can tell, their official chat site doesn't have a native audio -> audio model yet. I like to test this through homophones (e.g. record and record) and asking it to change its pitch or produce sounds.
39.
by potatoman22 11mo ago
I'd guess S&box is more an extension of Garry's Mod rather than a reaction to Unity
40.
by potatoman22 1y ago
After the Roman Republic, they switched to having an emperor. Jesus was crucified during this Roman empire. The kings of Rome were around 600 years before this. They meant the emperor, not the king.
41.
by potatoman22 1y ago
You got me into web dev. Thank you!
42.
by potatoman22 1y ago
I'm currently making a tycoon game with React, it's not bad for making some games. I use setInterval for a simple game loop along with a zustand store for the game logic. I'm keeping the game logic and state client-side for n
43.
by potatoman22 1y ago
The SpaCy creator has a good blog post on this https://explosion.ai/blog/against-llm-maximalism
44.
by potatoman22 1y ago
AI doing playtests is an idea I've been thinking about too. The question I can't quite answer is: how do you know the AI play-tester can predict what users find fun? How well does it represent the different kinds of users?
45.
by potatoman22 1y ago
Evidence for safe and fair AI systems is possible as long as you define what "safe" and "fair" mean for your usecase. Fairness might look like "no cohort has >5% higher false positive rate than another" and
46.
by potatoman22 1y ago
I'm afraid I don't quite understand your point. What line of thinking are you referencing? Also risk scores and algorithms have been used in medicine for over 50 years, so evaluating them isn't anything new.
47.
by potatoman22 1y ago
This is cool, but I’m a little skeptical. If Parachute uses AI agents to evaluate other models, who’s evaluating the AI agents? It’s hard to imagine it’s safe to entrust model validation and bias assessments to an automated system, especial
48.
by potatoman22 1y ago
Is anyone else sick of hearing "AI agent"? I can't fully explain the feeling, but it's nauseating. Y Combinator is probably the worst culprit in using it, especially on their YouTube channel.
49.
by potatoman22 1y ago
I mean you can call it AI/a luxury still. Tap water is nice, pathfinding algorithms are nice. It doesn't matter to me if people use "AI" whenever they want. Sometimes it's a little silly, but harmless. It's lik
50.
by potatoman22 1y ago
I don't think anyone is saying "you can't call facial recognition AI." I think their point is that laypeople tend to move the goalpost of what's considered AI.
51.
by potatoman22 1y ago
Some of the best applications of LLMs I've seen are for reducing bullshit. My goal for creating AI products is to let us act more like humans and less like oxen. I know it's idealistic, but I need to act with some goal.
52.
by potatoman22 1y ago
I feel like the hard part is flashing the new firmware
53.
by potatoman22 1y ago
I don't think that quote really supports coding and math being equivalent. To me, the quote provides an abstraction of math through a structuralist perspective. Language can also be viewed through this abstraction. I think coding could
54.
by potatoman22 1y ago
It'll be funny when someone's AI assistant buys them a different item than it said it would.
55.
by potatoman22 1y ago
Anyone know if the paper was published?
56.
by potatoman22 1y ago
Generating or augmenting data to train computer vision algorithms. I think a lot of defense problems have messy or low data
57.
by potatoman22 1y ago
It could be a man, but most relationships are heterosexual
58.
by potatoman22 1y ago
I can't tell if this is a joke
59.
by potatoman22 2y ago
What do you mean
60.
by potatoman22 2y ago
To contextualize, I think the tool in this article correctly diagnoses 97% benign cases as benign but misdiagnosis 22% of malignant cases.

More ›