I really hate how reddit has eaten the world. 15 - 25 years ago I had a stable of fun forums I would visit for my various technical and leisure interests and they were great. Now everything is getting slowly enveloped by the grey goo of AI slop. It's really hard to tell sometimes if you're interacting with a human there.

And making you log in so they can get your data is frustrating. I mean, if I'm not logged in I'm looking at the ads.

Reddit is addictive so it's taking a while but I'm trying to let it go. It just isn't worth it as it is actively being enshittified.

What are you talking about? You can buy quantum computers from a number of vendors today! You can buy one from IonQ, from Rigetti, IBM, Quantum Motion, and others. Do you mean a "utility scale" quantum computer? You should be more precise with your language.

So when are you sending me the money?

What strawman? I was quoting the article.

From TFA:

Not only was Claude trained on a lot of different C compilers, including the entirety of GCC, it still needed a golden model in order to finish.

If I claimed I had written my own C compiler from scratch and then this came out in an interview, I don't think I would be hired.

A lot of my colleagues love VSCode. The thing about vim is once you are proficient with it a lot of the commands become muscle memory. If you remember the days of vi vs emacs, I tried emacs a few times but kept coming back to vi because of the learning curve.

I learned vi in my first programming class 32 years ago and have been using it ever since. There are some things which are actually faster in vim than a GUI editor, but you have to have the muscle memory.

Good Lord. I feel called out. I just checked and I have 41 vim windows open, and I'm typing this comment on a mechanical keyboard. I also brushed this morning with a normal toothbrush.

Well, actually I don't have a record player or any vinyl (sold all my records in the early 1990s) so you aren't talking to me specifically.

This is awesome! Apple is paying through the nose for more crap I don't care about! Can't wait for the subscription fees to keep going up!

It's not the same Atari. It is the steaming remnants of Infogrames who wears the Atari skin like a zombie.

Atari first died in 1984, got resurrected immediately, then died again in 1996.

The current company likes to pretend they represent some kind of legacy of Atari, but it's bull.

You doing OK? Did I touch a nerve?

i have no idea if you actually used AI to write this reply, but it is hilarious that you use two of the "AI detector tropes" in this one sentence:

1. It isn't X, it's Y
2. em-dash (although in your case it's a dash).

I happen to live in San Francisco and I have never heard of Kyte. So, I'm not surprised they failed.

We have so many transportation startups founded here, Uber, Lyft, Getaround, Turo, Shift (RIP), and on and on. But never heard of Kyte!

That is why I said "analogous to a Ponzi scheme". In fact, the creation of a "distributed Ponzi scheme" may be the true innovation of crypto.

And haven't you got the memo, crypto isn't a currency anymore, hasn't been for a while. Now it's a "store of value" and an "investment".

Crypto is a trip. It generates no yield, has very little real-world use (beyond facilitating crime), and is a negative sum game.

Crypto requires real money to operate (a lot of custom chips are playing "guess the number" simultaneously), so the net amount of money coming OUT of crypto is necessarily less than the amount of money being invested in crypto.

In other words, the average return on crypto has to be less than 1. You have benefited personally, but that has come at the expense of others who came after (analogous to a Ponzi scheme). Crypto generates no value, so you just got taken out of the trade with new money.

It's all very interesting.

AGI is most certainly jargon. AI first began development in the 1950s and many goals of AI have been achieved. John McCarthy (of the Stanford AI Lab, and one of the field's founders) used to joke that once something was achieved by AI, it wasn't considered AI anymore.

Beating a chess master was once thought unthinkable but it happened almost 30 years ago. Similarly, many key AI focus areas, such as natural language processing and machine translation have made enormous strides just in the last two decades. Of course LLM's are AI. They are a powerful (but clearly overhyped) application of deep convolutional neural nets.

The "AGI" phrase is *more* a marketing term than "AI" is these days. In fact, the meaning of AGI is precisely what is expedient in the moment for whatever snake oil salesman is using it at the time.

Certainly not me.