smokel

22 ms·

391.
by smokel 2y ago
> They're afraid someone else might create something of value, in part using their API, without them getting their own beaks wet in the process. In most cases it's not about profit, but about having to invest serious amounts
392.
by smokel 2y ago
In the EU, this may soon be covered by the Data Act [1]. It'd be interesting to see if other continents will benefit as well, similar to how cookie banners are now a global thing. [1] https://digital-strategy.ec.europa.eu&#
393.
by smokel 2y ago
> games have clear objectives with obvious rewards. The real world, not so much. Tell that to the people here who are trying to turn their startup ideas into money.
394.
by smokel 2y ago
Yup, the technical details are described in this paper from early 2023. Seems like Nature takes its time. That used to be a good thing :) https://arxiv.org/abs/2301.04104
395.
by smokel 2y ago
You must be confused. This research is about reinforcement learning, not about large language models.
396.
by smokel 2y ago
This has probably little to do with skill being lost, and more with how little one gets paid to do this kind of work. Skilled people can get jobs in other fields that earn a lot more.
397.
by smokel 2y ago
This comment is a bit puzzling to me. The author of the podcast does indeed have a PhD in philosophy, but Paul Graham does not. The latter holds a PhD in computer science.
398.
by smokel 2y ago
> For most of history the question "What should one do?" got much the same answer everywhere This is so not true, that I'd like to point the author, and people who think in a similar vein, to a very enjoyable podcast on
399.
by smokel 2y ago
I, for one, love to think about having a lot of money that I can spend on fun things. I do, however, not love coming up with schemes all day to actually transfer that money from someone else's pocket to mine first.
400.
by smokel 2y ago
Note that there is a difference between wanting to have more money and wanting to make more money . The author seems to be lucky in being born with the latter trait, but most humans luckily don't. If you're in the other group,
401.
by smokel 2y ago
Here you go: https://en.wikipedia.org/wiki/Carceri_d%27invenzione
402.
by smokel 2y ago
I wonder what would happen in the event of a disaster where I have these 72 hours of supplies ready, but my friendly neighbor has not.
403.
by smokel 2y ago
> it's been an awesome force multiplier for small teams who need to productively pump out safe and ergonomic code. How much of this would be due to actual language features, and how much due to selection bias induced by talented a
404.
by smokel 2y ago
If you think it is implied that c stands for the first letter of the word causality , then reading the article should be enough to convince you otherwise. Einstein published his seminal works in German, and we'd be more likely to ha
405.
by smokel 2y ago
> So should we attempt such a thing, and if so, what would be our motivation? The thing is, this question needs not be answered satisfactorily before a single individual accrues enough means to make it possible. Apart from non-prolifer
406.
by smokel 2y ago
MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to LLMs. https://modelcontextprotocol.io/
407.
by smokel 2y ago
Code typically describes only the what and how. The why is usually left in people’s heads and, if you are lucky, in documentation.
408.
by smokel 2y ago
Perhaps give it some time before judging too harshly?
409.
by smokel 2y ago
Agreed, but I'm still very happy that some people try. I'm really not that much interested in the weather or listening to idle chit-chat, and for some reason most podcasts seem to focus on that.
410.
by smokel 2y ago
To see some contemporary "reverse oil painting on glass" works, see Jessica Frances Grégoire Lancaster's site. [1] https://www.jfglancaster.com/
411.
by smokel 2y ago
Also interesting in this context is the PyTorch Developer Podcast [1] by the same author. Very comforting to learn about PyTorch internals while doing the dishes. [1] https://pytorch-dev-podcast.simplecast.com/
412.
by smokel 2y ago
I'm interested to know if anyone is using fine-tuning to train a model on proprietary or in-house codebases and documentation. RAG solutions seem to have their limitations, and fine-tuning might be a more effective approach. How much e
413.
by smokel 2y ago
A simple text parser would probably be 10,000,000 times as fast. So the statement that this won't take much power at all, is a bit of an overstatement.
414.
by smokel 2y ago
> Absolute power corrupts absolutely. Might it be an option that correlation and causation are reversed here? Given the amount of criticsm a typical leader of a large company, or even a country, gets these days, it is no wonder that pe
415.
by smokel 2y ago
Don't underestimate how much effort it takes to apply a technique (LLMs) to a certain domain. Creating a good user interface, processing the code into appropriate embeddings for useful matching, and marketing are just three things that
416.
by smokel 2y ago
I don't think it is productive to compare a company to a nation state. Would you say the Finns are doing better as well, because Linus Torvalds was born there?
417.
by smokel 2y ago
"The Hitchhiker's Guide to the Galaxy" by Douglas Adams does a great job at being a timeless and priceless way to learn about the relativity of things.
418.
by smokel 2y ago
Reinforcement learning is hard to apply to real-world problems, but one cannot deny the success that a company such as OpenAI has.
419.
by smokel 2y ago
This looks like a neat idea, but I'm not too positive about it. This makes using computers even harder to explain to people who do not spend their entire day keeping up with the latest developments. They cannot form a mental image or
420.
by smokel 2y ago
It is interesting that you bring this to the attention, but I don't see why we should not trust or celebrate someone if they have views that you don't agree with. Edit: especially since I think your implied claim that Sutton would

More ›