staticman2

Claude is in (potentially) a price war and research and development arms race with OpenAI and Google, and potentially other labs, so I don't think it's necessarily a given to price it as the company with the "best devtool of all time".

Theoretically hedging would only work if the A.I. firms are not already priced for taking your job, right?

I have no opinion on whether they are.

> . However, the best law firms of 2029 will be using approximately the same tokens per employee as an A.I. lab of 2026, that much is clear.

This is as if you learned that Anthropic scanned 1000 books per employee this year to train their models and your conclusion was that the best law firm will be scanning 1000 books per employee by 2029.

Wikipedia tells me that "service industry" is a term that includes transportation companies and law firms.

So you are apparently now claiming my local train company or law firm will be using as many tokens per employee as an A.I. lab.

It's more accurate to say Anthropic intends to scam retail investors by claiming they will have the only A.I. worth anything.

Then if all goes well next year OpenAI will scam retail investors by claiming they will have the only A.I. worth anything.

If you have no theory of how to value a company any number looks good.

> If you fire that dev, who do you think will use Claude?

The remaining 10% of devs that I don't fire. How much software do I really need?

Anyway, I know there are some flaws in my argument, it was just a thought experiment.

I'm being sarcastic as to the specifics but didn't productivity go down 199k because the 200k engineer is now replaceable with a 1k subscription and will soon be fired?

Literally the output of the labor dropped to almost zero value?

I have not explored this but If I wanted Claude Code to make a spiritual successor to Punch Out I would first have it clone Punch Out to see if it was capable of doing this, then if it worked I'd start changing the assets and mechanics one idea at a time until I had something with more novelty behind it.

If you uploaded the audio via the web site or app and got terrible results I'd bet Argon wasn't the model doing the transcribing. The only way to know for sure is to try again with the API.

Or you failed to understand how the blog post using the eye of Sauron metaphor was cartoonishly misunderstanding the world we live in. Who can say?

It sounds to me like he wants you to believe the effective way to lobby for a ban on open source A.I. development is to use the same "classic" methods that someone in the 1950s used to lobby for coffee beans.

Therefore, as they are not acting like 1950s coffee bean sellers, they can't be doing regulatory capture.

That's the only possibly explanation why they don't lobby like coffee bean sellers.

These industries and eras are completely interchangeable.

It's not regulatory capture because look over here: I've invented a new term "classic regulatory capture" which involves only those historical incidents that are not like this.

This is by definition not "classic regulatory capture" a term that I have borrowed to cover things that are not this.

Whenever someone brings up "regulatory capture" I shall change the topic to "classic regulatory capture" which by definition cannot be this.

Very interesting quote.

> Google currently has the most generous usage limits on their $20 plan among US providers, so it's a pretty good deal.

Are you saying when you use Argon in antigravity it has generous token usage? I forget if I tried it with antigravity but my experience with using it as a chat bot was the usage limit was terrible.

Yeah this was hard to navigate on my phone but what I saw on my phone didn't look "stunning".

Every A.I. related post here has someone saying "I for one see no difference between a human being and an A.I." but this may be the first time I've seen this cliche extended to "How bizarre to meet a human being that thinks humans are not identical to A.I."

As I understood it the less wrong people think the A.I. will likely rapidly cause human extinction.

I'm not familiar with/ wasn't familiar with any "economic" phases of the less wrong lore.

My comment had nothing to do with the economy. I was finding humor at the visual of a guy with authority over an actual existential threat (nuclear arsenal) being told the lesswrong lore about graphic cards in data centers being far more dangerous to humanity, and presumably, thinking the less wrong guy has something wrong with him.

I can't imagine why someone with control of a nuclear arsenal wasn't wetting his pants over the existential risk of graphic cards.

Why do you expect it to crash before the OpenAI IPO?

> Anthropic thinks it is.

Does that have to be true? I thought IPOs were about selling the company.

Never mind the fact that companies do not think and this is like saying "My lemonade stand thinks it's going to have good sales this quarter."

I will note William Gibson definetly wasn't in any traditional sense a tech enthusiast.

In Novemember 1994 he said:

"I don't have an e-mail address, I don't even have a modem. As much as I admire the Internet I suffer literally agoraphobia, which in it's original sense means a fear of the marketplace. I do not want to receive three hundred e-mail messages per week from strangers wanting to communicate with me. If only because I'd be tempted to open them all and look at them. And there goes, you know, half the time that I have to write. I mean, the amount of physical mail and other communications I get these days is already swamping me. "

https://josefsson.net/gibson/gibson4

Wouldn't a delay due to synthetic COT generation look pretty much the same as a delay due to insufficient compute for all customers?

If you only have to look at revenue to determine the value of a company stock picking must be very easy for you.

Calculate the price to revenue ratio of all companies and select whichever ones have the best ratio.

Your question is vague and you didn't explain your point of confusion but consider the fact that securities laws are not fungible between nations.

I agree we shouldn't believe Anthropic but if you wanted to scam people into thinking they weren't using Claude wouldn't you simply generate synthetic COT?

> What I don't understand is, why would Google allow such direct competition to their own AI subscription?

They are much better off with you using "Google docs" then if Claude or someone else were to launch "Claude docs".

Suppose the big 2 labs didn't think the expense was "worth it."

Wouldn't they keep spending anyway because they need to hype things up before IPO?

If Claude's custom solution is less secure than LiteLLM there will be no news report informing you of a vulnerability.

If you stick with older LiteLLM versions you won't get proper support for new LLM releases, correct?