giveaccountpls

My friend, please read on what functional languages typically are, vs what imperative languages typically are, before sharing your thoughts. It's pointless to engage your post as you clearly do not understand the basic jargon.

still waiting for a better result than "a anon-optimizing compiler where progress is impossible": https://www.anthropic.com/engineering/building-c-compiler

Funnily enough, the models learned from GGC and LLVM, yet, cannot produce something half-decent.

I've read the wikipidea page of MIRI, and it does not mention any achievements; just that it was founded "with the purpose of accelerating the development of artificial intelligence". Did they advance the field of AI alignment at all? Being concerned with a problem does not make one an expert.

For the rest, is he just an AI influencer? One that predated his topic of interest, but still just an influencer?

What are his actual credentials? What did he achieve? Isn't he just some random dude on the internet? Why do people care what he says?

You're ignoring the point of the slides: Tao thinks math is about understanding, not results. AI brings results. He is pondering how to use AI to increase understanding.

FWIW, as a non-drug-discovery guy, I understood his example, I believe, as intended.

Python is compiled (fun fact: it has stronger typing than c++). Java is also compiled, so is C#. All of them require a runtime. Your comment is silly.

Sorry, if your job title implies even a smidgen of security responsibility, you deserve to be fired for "123456" as your password. The person who was an administrator would fall under this label. Besides, "If productivity is not your goal, then security is not my goal." is what results in draconic security measures, because employees can't be trusted AT ALL. I really don't understand your comment.

> and therefore see no reason to try and coexist with it

This is my main point of disagreement. This is not their decision to make.

> The authors can’t assume you’ve read the manual or understand what the program was designed to do

I think this is correct, but, the principle of least surprise should be paramount here. If the program states something like "migrate all vim-related files to neovim(This will lose all metadata when opened by vim)? y/[N]", I'd concede that neovim wiping data if the user chose Y as acceptable. Yet, they neither display the prompt, nor keep existing data.

Because I can judge the return of for-profit companies. At least the ones selling a product, especially a physical one. The ones that don't sell a gauge-able product should, to my judgement, ascribe to the level of transparency that charities should provide.

Disagree with the notion of taste in this article. Especially, since it mentions taste in the context of FE. Taste, funnily enough, of people with good taste, is extremely valuable. When reviewing very large diffs, the fastest, and therefore, the most valuable, approach is to take a glance and say "this doesn't look right, it should look like X".

Besides that, I agree that coding isn't solved. In fact, I think coding is the new "hard things are easy, easy things are hard". Models, yes the newest and most expensive ones included, have atrocious taste. They will work very hard to cover every edge case they can conceive, without realizing the solution is to simplify.

I will entertain the notion that models solved coding when they can implement that gcc alternative (https://www.anthropic.com/engineering/building-c-compiler) and actually produce a half-working compiler. They'll get extra points for not reusing the test suite. Even more points for models that have somehow not trained on gcc's codebase.

Finally, remember that a compiler is something of a best case scenario for models. They are, UB withstanding, pure transformations. The models can write millions of tests that will run in a reasonable time. Real-life systems are usually not as testable, and are a much harder challenge for LLMs.

For an organization begging so aggressively for donations, Wikimedia got $150M in donations during July 1, 2020—June 30, 2021. They spend 2M on hosting, 67M on salaries, with 9M of it going to management, and pocketed 50M. Moreover, it's not clear how the 50M they spent on Programs is split between:

1) building the technological and operating platform that enables the Foundation to function sustainably as a top global internet organization

(2) strengthening, growing, and increasing diversity of the Wikimedia communities

(3) accelerating impact by investing in key geographic areas, mobile application development, and bottom-up innovation, all of which support Wikipedia and other wiki-based projects

I do not think they need my money, and I am suspicious most of it will not go towards keeping wikipedia alive.

I expect the people asking for money to do good, to show that they are doing the good they said they'd do. I am essentially giving them money with almost no liability, I want them to show me that they are not wasting/splurging it. Perhaps I'm too harsh, but given the prevalence of grifters/scammers these days, I find non-transparent charities trustworthy.

Given wikimedia's ranking on this website, I am not so certain it's worth engaging. Obviously the "did not share info therefore guilty" mantra is somewhat true for charities. But "did share info therefore not guilty" is even more silly.

The guy you're arguing with made the mistake of switching to metaphors. The result is several back and forth posts, and neither mutual understanding nor useful discussion.

To avoid metaphors: There's an implicit assumption that when you install a piece of software, it will not fuck up your data/system without your consent. No, installing it is not consent, neither is using it. Neovim does exactly that. No amount of food analogies or license quotes will change that fact.

Neovim authors should've just used a separate file, just like they're happy to use separate format and files for the configuration. Instead, they've decided to be dickheads, and interfere with vim's file, even after the issue being pointed out to them.

I only disagree with looking down on people who dislike poems that don't rhyme. Rhyming is another way to demonstrate effort, a thing that the author claims is thick. Not to say I dislike all non-rhyming poems, but it is just less impressive to capture an idea or feeling without rhyming, than it is with.

One of the signs of a declining band is the discarding of rhyming. This is usually accompanied by uninspired lyrics. The result is often something only for the die-hards to enjoy.

Yes. The internal services should be doing stuff in bulk and not require too much parallelism. That leaves you with the number of concurrent users, which in b2b apps can be quite low.

> Giving up your own personal stake in the process is literally what brings something as beautiful as Phish's music into existence

Unless Page McConnell is creating his music with AI these days, I don't think this point holds.

> neurodivergent people in our ecosystem who would trip up on that question just because of how it's framed

If anything, it's neurodivergent interviewers. If I insisted on a different design I'd either ask a question that's not solved by "just use postgres" or follow up with "ok, that would work, but what if <something that would prevent postgres from working>". Just failing a candidate for a correct answer is a prime example of why interviewing is so bad.