Certhas

11 ms·

181.
by Certhas 11mo ago
Non-political almost always means "accept the social status quo I am used to". I think there is a reasonable argument that the default for a community with technical goals should be to accept social status quo conventions unless t
182.
by Certhas 1y ago
It isn't working well by what standard?
183.
by Certhas 1y ago
The parliament approves and dismisses the commission. In the last cycles the candidate who led the party who won the parliamentary elections became head of commission. So this is just wrong. The EU parliament has more power than US Congress
184.
by Certhas 1y ago
So this is typical of criticism of the EU democratic structure: It's just factually wrong. The EU Parliament can dismiss the commission. From Wikipedia: "The Parliament also has the power to censure the Commission by a two-thirds
185.
by Certhas 1y ago
The EU is a parliamentary democracy. Von Der Leyen was proposed by the democratically elected heads of the member states. She was approved by the democratically elected parliament. The chancellor in Germany is also not directly elected by m
186.
by Certhas 1y ago
In another post discussion, someone suggested git as an alternative to overleaf, a Google Docs for latex... I guess there are plenty of people with blind spots for things that are technically possible, and usabel to experts, and UI that act
187.
by Certhas 1y ago
These are great suggestions, but did I understand that suggest git as an alternative to overleaf? That's... not at all reasonable, as well as completely missing what problems overleaf solves for which people. Overleaf ist git und the h
188.
by Certhas 1y ago
AMD has had unified memory for ages in HPC and for a while now in the Strix Halo systems. I haven't had the chance to play with one yet, but I have high hopes for some of our complex simulation workloads.
189.
by Certhas 1y ago
I am super intrigued by such world models. But at the same time it's important to understand where they are at. They are celebrating the achievement of keeping the world mostly consistent for 60 seconds, and this is 720p at 24fps. I th
190.
by Certhas 1y ago
Well you missed the point. You could call it prompt adherence. I need veo to generate the next frame in a few milliseconds, and correctly represent the position of all the cars in the scene (reacting to player input) reliably to very high a
191.
by Certhas 1y ago
This _might_ be true, but it's utterly absurd to claim this is a certainty. The images rendered in a game need to accurately represent a very complex world state. Do we have any examples of Transformer based models doing something in t
192.
by Certhas 1y ago
We have successfully banned Cigarette ads. Not saying there aren't things to sort out, and lines to be drawn, but it's by no means impossible.
193.
by Certhas 1y ago
I think this is still too extreme. A machine that cuts and preps food at the same level as a 25th percentile person _being paid to do so_, while also being significantly cheaper would presumably be highly relevant.
194.
by Certhas 1y ago
I haven't kept track of numba in recent years. But there is a clear path to translate more and more scikit-learn to mojo, bypassing the python interpreter entirely. And then things become much more composable in a way that numba can&#x
195.
by Certhas 1y ago
As others said: you can. And what's the alternative? Not being able to read the data or never being able to evolve the encoding.
196.
by Certhas 1y ago
Is this relevant in practice? Say I go to a website to download some data, but a malicious actor has injected an evil decoder (that does what exactly?). They could just have injected the wasm into the website I am visiting to get the data!
197.
by Certhas 1y ago
If you have a native decoder you don't need to run WASM. It's literally in the abstract.
198.
by Certhas 1y ago
Agreed on your last point. It's fairly obvious I think that we won't factor 1024 bit numbers by any direct evolution of the techniques used in smaller circuits. The hope is that you reach a threshold (error correction) where a com
199.
by Certhas 1y ago
It even says so right in the abstract: "Embedding the decoders in each file requires minimal storage (kilobytes) and ensures compatibility on any platform in case native decoders are unavailable." The idea that software I write to
200.
by Certhas 1y ago
So the argument is that because this doesn't solve the challenge of ordering all possible strings by semantic meaning it should not be used? Even though it increases the match between semantic meaning and string sorting in many importa
201.
by Certhas 1y ago
The sort rules are simple (1). Treat any consecutive sequence of digits as a number when sorting. So for example version numbers (which must be massively more common than decimals in filenames) work correctly, and 5.9 is indeed smaller than
202.
by Certhas 1y ago
Agreed.What's more, the idea that people learn to put leading zeros is wrong and impractical, unless you know in advance how many digits you need. When you go from version 5.9.17 to 5.10.0 you don't go back and relabel every exist
203.
by Certhas 1y ago
Actually no. 50×(16×6 + 32) = 50×(16× 8) works I think. Every 50 moves, move a pawn or capture. There are 16 pawns. Each pawn can be moved 6 times, so there are 16×6 pawn moves available. In addition there are 32 captures available.
204.
by Certhas 1y ago
PostScript is turing complete. SVG almost got the ability to open raw network sockets. Word files were (are?) a common platform for distributing malware.
205.
by Certhas 1y ago
Obviously^1. But it illustrates the broader point of the article, even if for the concrete problem even better choices are available. 1) if you have studied these things in depth. Which many/most users of solver packages have not.
206.
by Certhas 1y ago
It hasn't. Multipolar world, expertise exists everywhere. But user-facing innovation is coming from the US. No EU Apple, Google, Amazon. And infrastructure R&D in China is unprecedented. They are reaping a multi-decadal investment
207.
by Certhas 1y ago
People really don't appreciate what is possible in infinite (or more precisely: arbitrarily high) dimensional spaces.
208.
by Certhas 1y ago
I know that if you go large enough you can do any finite computation using only fixed transition probabilities. This is a trivial observation. To repeat what I posted elsewhere in this thread: Take a finite tape Turing machine with N states
209.
by Certhas 1y ago
Take a finite tape Turing machine with N states and tape length T and N^T total possible tape states. Now consider that you have a probability for each state instead of a definite state. The transitions of the Turing machine induce transiti
210.
by Certhas 1y ago
We didn't but somebody did so it's possible so probabilistic dynamics in high enough dimensions can do it. We don't understand what LLMs are doing. You can't go from understanding what a transformer is to understanding w

More ›