willj

5 ms·

1.
by willj 3mo ago
How can you tell if someone is a polars fan? Don’t worry, they’ll tell you. :)
2.
by willj 4mo ago
Fair point. Though I don’t think time without money is really leisure time :)
3.
by willj 4mo ago
It’s too bad that, yet again, instead of the productivity gains leading to shorter work weeks, the benefits accrue to the companies. Just once I’d like to see productivity gains lead to more leisure time, not higher expectation.
4.
by willj 7mo ago
The temperature parameters largely went away when we moved towards reasoning models, which output lots of reasoning tokens before you get to the actual output tokens. I don’t know if it was found that reasoning works better with a higher te
5.
by willj 7mo ago
If you’re using a model from a provider (not one that you’re hosting locally), greedy decoding via temperature = 0 does not guarantee determinism. A temperature of 0 doesn’t result in the same responses every time, in part due to floating
6.
by willj 7mo ago
A temperature of 0 doesn’t result in the same responses every time, in part due to floating-point precision and in part to to lack of batch invariance [1] [1] https://thinkingmachines.ai/blog/defeating-nondeterminism-in
7.
by willj 8mo ago
Thanks! That makes sense. I suppose this requires commit messages or PRs to indicate code was AI-generated vs. not, or to assume that commits after a certain time period were all from AI coding. It’d be an interesting analysis. Maybe there’
8.
by willj 8mo ago
100%. This is what I posted about on Hacker News ([1] where it got no traction) and Reddit [2] (where it led to a discussion but then got deleted by a mod). [1] https://news.ycombinator.com/item?id=46705588 [2] https:&#x2F
9.
Ask HN: What are the metrics for "AI-generated technical debt"?
3 points by willj 8mo ago|2 comments
10.
by willj 8mo ago
Can you say more about the approach you take for summarization? Are the papers short enough that you just put the whole thing in the context window of the model you’re using, or do you do anything fancy? I’ve tried out various summarization
11.
by willj 8mo ago
Thanks! I got some initial ideas from Nano Banana, actually, but then spent a while iterating on different layouts myself.
12.
by willj 8mo ago
This is something I built over the holidays to support people having a hard time with the short days and early sunsets: https://sunshineoptimist.com . For the past several years I would look up the day lengths and sunset times fo
13.
Show HN: Sunshine Optimist: Optimistic takes on daylight and sunset times(sunshineoptimist.com)
1 points by willj 8mo ago|0 comments
14.
by willj 8mo ago
I think the models are so big that they can’t keep many old versions around because they would take away from the available GPUs they use to serve the latest models, and thereby reduce overall throughput. So they phase out older models over
15.
by willj 9mo ago
This reminds me a bit of using LLM frameworks like langchain, Haystack, etc., especially if you’re only using them for the chat completions or responses APIs and not doing anything fancy.
16.
by willj 11mo ago
DOOMscroll[1] for sure! I still play it since hearing about it on HN. [1] https://ironicsans.ghost.io/doomscrolling-the-game/
17.
by willj 2y ago
I think this ignores that the monopolies have the power to buy up any new competitors, or to drive them out of business using monopoly power. Regulatory hurdles are only one tool that (can) benefit monopolies.
18.
by willj 2y ago
I think that’s different. AlphaGo is using reinforcement learning in a context in which there is a clear evaluation function— did a strategy lead to a win or loss.
19.
by willj 2y ago
Relatedly, the OCR component relies on PyMuPDF, which has a license that requires releasing source code, which isn’t possible for most commercial applications. Is there any plan to move away from PyMuPDF, or is there a way to use an alterna
20.
by willj 3y ago
I’d argue Bitcoin is Obscene Energy Demand.
21.
by willj 3y ago
Where are the happy offices these days? Which companies are the new “Google” who people are very eager to work for?
22.
by willj 3y ago
That makes me think of this: https://nohello.net/en/
23.
by willj 4y ago
I like this idea, but isn’t this a recipe for only ever doing the urgent stuff, not the not-urgent-but-important stuff? For example, if your list had “read 1 chapter of SICP” on it, you might never get to it.
24.
by willj 5y ago
One thing with fastai that annoyed me when I built a project with it was that the v1 and v2 APIs are totally different, and if you google or search stackoverflow for help with something, I found it more likely to stumble on answers for the
25.
by willj 5y ago
I’m not sure if you’re aware of this option already, but Google Colab sounds ideal for this situation. People get a copy of your notebook, which they can run, edit, and try new things, and all the computational resources are free. It even o
26.
by willj 5y ago
I do the same, but I think there are an increasing number of astroturf campaigns taking advantage of that, too.
27.
by willj 5y ago
Can you imagine if someone had kept their crypto wallet in this thing?
28.
by willj 5y ago
Makes me wonder how the engineers will fix this if they can’t visit Stack Overflow :)
29.
by willj 5y ago
Taxing a quantity is way simpler to implement, and maybe more effective, than making a huge list of items to tax, thereby creating loopholes.
30.
by willj 5y ago
Isn’t this the same as “pay with PayPal/Amazon pay/etc”?

More ›