> Also, remember that NATO pushed Ukraine to start having civilians manufacture drones in their homes.

Dude.. Take your meds.

That may be true, but I have seen no evidence this is a routine thing (or a thing at all) like it seems to be for Russia. So far no Ukrainians have been indicted for such war crimes, whereas multiple Russians have been.

There's also the optics of it. If it were to be discovered that Ukraine's e.g. intentionally leveling residential areas around Moscow, support would instantly dry out. It makes no sense for Ukraine to waste ammunition on targets that hold no military value. It would accomplish nothing _and_ they'd lose the support of the rest of Europe.

>> Collateral damage is sadly unavoidable.
>Remember that when you see the next Ukrainian news that Russia bombed a random civilian building ... maybe that were not the target, that surely is not disclosed to not help the other side, but simply a collateral damage

I have no doubt that some of Russia's damage to civilian structures is accidental. I also have no doubt that the majority of it is intentional. There are currently 6 officials wanted by the ICC for war crimes in the Russia-Ukraine war, all of them Russian.

>so who do you know that is "intentional bombardment"?

See above.

I have no idea what you're trying to comment on. Did you click reply on the wrong post?

Striking production facilities, military personnel and supply lines in an invaders own territory is absolutely fair game. Collateral damage is sadly unavoidable. Not all drones will reach their targets, some targets will be based on invalid intel, and civilians might come in harms way as a result. That's a far cry from Russia's intentional bombardment of civilians and related infrastructure, though. One is a war crime, the other is not.

Considering reports now are that Russia's finally on the back foot, it seems Ukraine is doing what it needs to do. If they were wasting munitions on apartment complexes instead of strategic targets, that would not be the case.

A pull request is not "downloading the source code," it's "submitting a suggestion for a change to the source code." It's someone requesting the project "pull" their changes into the project.

..of removing actual intelligence in order to replace it with "intelligence-native" AI, which is all A and no I.

Apparently Block deals with financial transactions. I see no reason to worry at all. /s

Take code written in a time when people writing code actually knew what they were doing, because it was so bloody annoying to do that only people really interested in doing it well were actually doing it.

Let's make it code that handles money transfers, an area even the most ignorant would likely agree requires code of high quality.

Let an LLM rewrite that in a modern language, using training data from a whole lot of really low quality code scraped off the internet, and with no one competent in the original code available to ensure the result is anywhere near sensible. Sounds like a great idea.

Predicting the future isn't easy.

The .com bubble was pretty much "Oh hey, computers have been around for a while and they're great, now the internet is arriving so let's make sure we don't miss out. Let's pump all of the money into any company that has a web page or might consider making one, in case they conquer the internet."

LLMs feel to me more like, "Oh hey, someone added an LCD screen to a toaster and they claim next year they'll make the airline industry irrelevant, let's put all of the money into toaster factories!"

That doesn't mean LLMs will have no place where they're useful. But, barring a new breakthrough, it's hard to see how what is in essence a "pick next word" algorithm might turn out e.g. a new Google.

But, as mentioned, predicting the future isn't easy. Perhaps I'll look like a luddite in a few years time. Then again, my track record of greeting 3D movies with "meh" every time they roll back into fashion, is pretty darn good. :p

There are ten links in this summary. What is the method by which you decide the first sentence of the third link is the one that makes you decide whether to keep reading or not?

This is starting to look like bad faith argumentation to me. Remember that the premise here was using LLMs as assistants for code review, not code production. Decent coding assistant LLMs that can be run locally already exists, and have for years. They're not as good as full size models running in data centers, of course, but they're also not useless. As long as they are able to catch _some_ security issues, at an acceptable signal to noise ratio, they offer value in freeing up reviewer brain bandwidth to focus on other issues.

The problem arises when people buy into the "these things are so brilliant" hype and figure they're a silver bullet for everything. They're definitely not. They have a very niche usage area, but within that they can be useful. And with models now existing on millions of computers around the world, for languages that are not going out of fashion any time soon, and with people making old mistakes all the time, I don't see that vanishing anytime soon even if big tech folds. We'll just have to wait to see how it all pans out.

Based on the summary, to me the use of "upstream" indicates "maintainers". As in the people responsible for approving and merging.

I'm not sure the business numbers are all that important when it comes to code. We already have them trained on _a lot_ of code, and since they're more focused they can be smaller without being useless compared to the full size ones. If we can run them locally on a single GPU, it doesn't go away when the bubble pops and the big players stop throwing away money.

As with any tool, they need to be used where they actually offer value. Which is definitely not to architect solutions, but to sanity check smaller chunks of code. Assuming false positives can be kept to a minimum, that is real value even if all they do is save some time in catching a limited set of issues.

I've heard him talk about that as well, so he seems well aware of it. Of course, that doesn't automatically mean he'll be brilliant at actually doing it himself. It's one thing to talk about other projects in hindsight, something quite different when doing it yourself.

Personally, I'm cautiously optimistic.

He's stated they did not listen to his input on that show, so unless he's lying about that it shouldn't be much of an indicator. Of course, it could also be his input was even worse than the choices they went with. I find that unlikely, but it's at least a possibility.