choeger

24 ms·

241.
by choeger 3y ago
Dear Liberty Media, please have a look at such a tool and maybe reconsider your addition of defective by design (aka DRM) technology to f1tv streams. It broke on three (!) different playback devices I own. thank you, a valuable customer
242.
by choeger 3y ago
Simply put, OCaml is very principled. There is a good reason for pretty much every design decision. The only thing one could argue is the lack of type classes, i.e., overloading. (Calling it "interfaces" is somewhat wrong.)
243.
by choeger 3y ago
presumably still O(n2) in theory, but not for practical cases. I think that anything reolacing attention will suffer quadratic growth for some pathological examples. maybe if we have a better understanding of the data we could give a better
244.
by choeger 3y ago
That's refinement-style programming from a novel angel, but still clearly refinement-style.
245.
by choeger 4y ago
Cool project. I think we really need some free software sensor/controller libraries/frameworks for solar setups. The reason is that here in Germany, and presumably elsewhere in Europe, these "balcony powerplants" are tak
246.
by choeger 4y ago
Nice try, I am not sure if someone can escape the user: prefix. But if that model reads and summarizes an email, your defense is pointless, because the email's content will be behind it. Basically, every plugin must somehow be able to
247.
by choeger 4y ago
In what sense is TCP tomorrow different from disallowing 3rd party cookies today? On a side note: Isn't it odd, how the composition of different technologies (here HTTP and DNS) make it unintuitively complex to implement a so intuitive
248.
by choeger 4y ago
I think that there will be a new successful browser one day - and it will be disruptive. But it needs two properties: 1. A unique use case or feature that cannot be easily implemented in the existing browsers. Something that breaks the cu
249.
by choeger 4y ago
I have yet to encounter a single valuable (in terms of hard currency) use case for GPT besides automating publishing. Yes, you can generate text automatically now in relatively high quality. But that's a business that was struggling be
250.
by choeger 4y ago
You have obviously no clue what "theory" means in that context. Playing with LLMs has so far demonstrated very little. Yes, these language models are subjectively impressive, but that's exactly what they have been designed to
251.
by choeger 4y ago
Impossible. In order to fix a bug, you need to do abstract interpretation or something equivalent. Language models simply can't do that. And even the best bug report will not tell you what you need to do to fix it. Of course, there are
252.
by choeger 4y ago
I tend to agree with the "bullshitting" part. But that gives us a very useful tool to estimate future applications of LLMS. Will a LLM ever program on its own? Certainly not. It gets semantics wrong so often, it's not even fu
253.
by choeger 4y ago
Is that the new hot take after "costs are too high because scale is too small"? There are many issues with the economics of nuclear power but too much safety isn't one of them. The Chernobyl comparison conveniently ignores th
254.
by choeger 4y ago
I wonder if we should go back to ontologies now. It should be fairly easy to have a LLM generate entries for any ontology. Then we need to check them for truth, obviously, but that can be parralized and potentially automated. Say, I'd
255.
by choeger 4y ago
Can someone ELI5 how linear attention works? Isn't that impossible in principle?
256.
by choeger 4y ago
I have to ask. Why is it "white"-board vs. "chalk"-board? Back in elementary school, I learned to call the former "black"-board.
257.
by choeger 4y ago
> There is no amount of beef that is good for the environment. That's a tautology. In that strict sense, there's no amount of you that's good for the environment, either. Unless we all commit suicide, we have to accept t
258.
by choeger 4y ago
I often wonder, how much double counting goes into these stats. Of course, dairy currently relies on fossile energy often. But how much of it is attributed elsewhere as well? Diesel, for instance, is often also attributed to big oil compani
259.
by choeger 4y ago
This. There needs to be a pyramid in top of the food pyramid. Beef is not bad per se, it just shouldn't be your weekly meat.
260.
by choeger 4y ago
Absolutely. I had a 25% discount offer for the same flight when booking through the American site vs. the European one. Unfortunately, they also asked for an invoice address in the states. So I opted out of that airline altogether. Just wai
261.
by choeger 4y ago
If your make targets don't have dependencies you are probably using the wrong tool for the job.
262.
by choeger 4y ago
Agile as you observe it ist just double speak for micromanagement in my experience. What we experience is a caste of highly paid managers that a) don't understand the work their subordinates are doing, b) don't trust them and thus
263.
by choeger 4y ago
On top of that, I think we might have reached peak training data. Or rather, peak human-generated training data. From here on AI will produce much more data than humans. Consequently, it should become ever harder to train a model on somethi
264.
by choeger 4y ago
Well, we could fingerprint the fingerprint detection code ...
265.
by choeger 4y ago
What happens if you twist that joke into something that doesn't make sense? Like having the bartender ask "can I bring anyone of you a drink?"
266.
by choeger 4y ago
So it definitely does a little bit more than just dumping math queries to a CAS. Intriguing.
267.
by choeger 4y ago
> His images have been bootlegged endlessly by Asian companies placing them on t-shirts and selling them for $5 on Etsy and every other marketplace. And why not, really? How does that affect your brother? Why should your brother have a s
268.
by choeger 4y ago
How do these models handle renaming? E.g., "Let foo be the function that calculates the sinus of its first argument. Let bar be the circumference of a circle with radius one. What is the value of foo applied to bar?
269.
by choeger 4y ago
I did a ton of parsing back in the day. LL, LR, LALR, in various forms with various lookahead, and in several languages. I once even did an incremental LR parser generator. I used, but never fully implemented, Pratt and Early parsers. I s
270.
by choeger 4y ago
Correct me if I am wrong, but that's actually a smart move, because the securities cover these liabilities, no? AFAIK, SVB has securities that are currently rated below their nominal value by the market. But these securities will rea

More ›