atemerev

35 ms·

271.
by atemerev 1y ago
Thank you for your kind words! Re: the woman in France, sadly, there are many people like her. Many of them _had_ obtained the citizenship of their host countries.
272.
by atemerev 1y ago
Oh I would be the most happy person to do it. The writing was on the wall since 2008, and I am painfully aware of it. However, my host country is Switzerland, where if you are not already a EU citizen, you have to live 10 years basically wo
273.
by atemerev 1y ago
Universal quantifiers are not useful in this discussion.
274.
by atemerev 1y ago
Well, American society supports Trump, right? and Trump wants to invade Canada and Denmark. No, "Russian society" is not a monolith supporting the war of aggression. Just like American society is not all like their government offi
275.
by atemerev 1y ago
Russia was not a member of CERN, so most Russian physicists worked there on (fake-ish, for documents only) exchange programs with JINR Russia. Most of them were living in France or Switzerland for many years. Given the opportunity, every si
276.
by atemerev 1y ago
Oh well. When Russian scientists were escaping the horrible realities of Putin's regime, had someone in Europe attempted to "lure" them? No, they were fired by hundreds [1] and students were not allowed anymore [2]. Not count
277.
by atemerev 1y ago
170 days of Duolingo is not enough. It sort of works... but yes, it takes a few years of daily focused work.
278.
by atemerev 1y ago
So basically Mycoria IP addresses/keys leak information about your geographical location?
279.
by atemerev 1y ago
Then you are free to not use their product. My privacy is indeed differential. I am willing to give them the information on my coding patterns and even non-commercial code for a free license, if it is not linked to my identity. This is a fa
280.
by atemerev 1y ago
People entrusted with investing billions of dollars are very much not rational, as we can observe every day. Additionally, network effects and experience input and VC concentration in Silicon Valley are very real and very rational. For VCs,
281.
by atemerev 1y ago
Their telemetry promises not to collect private data. Yes, your code will probably used for training their models. But so it would be if you publish it on GitHub.
282.
by atemerev 1y ago
Well, that's the difference between being in Silicon Valley and not being there.
283.
by atemerev 1y ago
We have to do arbitrary insertions/deletions from the middle, many of them. I think it is more like BTreeMap, but we need either sorting direction or rev(), and there were some problems with both approaches I tried to solve, but eventu
284.
by atemerev 1y ago
And it will look like this: https://rust-unofficial.github.io/too-many-lists/sixth-final... (filled with boilerplate, strange Rust idioms, borrow_unchecked, phantomdata, and you still have to manage lifetimes annotatio
285.
by atemerev 1y ago
I have read the "too many linked lists" story and I think the other commenters here are right; the less pointers the better. Even with unsafe, there's just too much ceremony.
286.
by atemerev 1y ago
Right. I attempted using Rust for trading-related code as well. However, I failed to write a dynamically linked always sorted order book where you can splice orders in the middle. It is just too dynamic for Rust. Borrow checker killed me. A
287.
by atemerev 1y ago
The block is on the Russian side (in most cases), but not all Western sites are blocked (Hacker News is working). Most of Russians know how to use VPNs, though it is extremely inconvenient.
288.
by atemerev 1y ago
For the record, I don't think that this particular event is a Russian act of sabotage. But Russia is an aggressive authoritarian state that was already caught for (smaller) acts of sabotage in EU, some of them quite dangerous. Why they
289.
by atemerev 1y ago
Parts of EU.
290.
by atemerev 1y ago
Depends on intensity. Microgram quantities of plutonium should be generally safe (unlike, say, microgram quantities of polonium). Not all alpha emitters are created the same.
291.
by atemerev 2y ago
Nice. I need to batch process my entire database of addresses (100M entries), so only local Nominatim will work.
292.
by atemerev 2y ago
The evil is now caricature, like a villain in the most cliché Hollywood movie ever, taking pride in how evil it is; and there are _still_ enough supporters of evil for it to be viable. I don't know how it even works.
293.
by atemerev 2y ago
There is no OpenAI model better than R1, reasoning or not (as confirmed by the same Aider benchmark; non-coding tests are less objective, but I think it still holds). With Gemini (current SOTA) and Sonnet (great potential, but tends to over
294.
by atemerev 2y ago
Yes, confirmed by citing Aider benchmarks: https://openai.com/index/gpt-4-1/ Which means that these models are _absolutely_ not SOTA, and Gemini 2.5 pro is much better, and Sonnet is better, and even R1 is better.
295.
by atemerev 2y ago
But this is not "system prompt"
296.
by atemerev 2y ago
This is what Llama4 replies (locally running, no system prompt): "I'm Llama, a Meta-designed model here to adapt to your conversational style. Whether you need quick answers, deep dives into ideas, or just want to vent, joke or br
297.
by atemerev 2y ago
And yet, if you ask Claude, Llama, or xAI who they are, they answer correctly. Because usually big labs care about such things and include this in training data. So, not a proof, but some evidence.
298.
by atemerev 2y ago
Most of these tools are focused on blocking apps. I read everything through browser.
299.
by atemerev 2y ago
10 million records is a toy dataset. Usually, you can fit it in memory on a laptop. There are open large(-ish) text datasets like full Wikipedia or pre-2022 Reddit comments, that would work much better for benchmarking.
300.
by atemerev 2y ago
"In fact, if you pitched this pathway to a VC, I’m sure they’d ghost you quicker than their last Hinge date." — oh they will, fr

More ›