What was their original mission?

My sense is that startup mission statements are ~meaningless. Builders try to build great things that lots of other people will find valuable.

>What was their original mission?

Beat OpenAI. The Founders came from OpenAI so there was obviously some disagreement about the direction there or they simply wanted more control.

Google will beat OpenAI on general AI. Anthropic have wisely leaned into the thing they're best at, which is coding.

Friendly AI

To maximise shareholder value.

Technically speaking, it's to maximize shareholder value while serving the public interest. They're a public benefit corporation.

OpenAI has "open" in their name but is closed off to public access and input

Google used to have a motto "don't be evil"

Who enforces the definition of language? Who demands compliance?

Soon as we go down the path of policing and insistence on one true dogma, we veer into religious holy war type behavior.

Obsession with semantics of syntax is a sort of theism even if the syntax and semantics do not refer to the commonly accepted tropes of a specific religion.

In this case, it has to do with how they're classified under Delaware law as a corporate entity.

I'm not a lawyer (I don't even play one on TV, damn you Odenkirk) so I can't tell you what that means as far as case law for companies getting punished for behaving badly, but in this case, there is supposedly some sort of legal backing for the classification.

Laws are words with zero meaning if they go unenforced.

Politicians are not interested in assuring such.

Public is busy arguing semantics online; they are not interested in assuring such.

In other words, you didn’t know they’re a PBC, and you don’t know what that means?

I know PBC; legal semantics that dictate various financial terms. I also know the specifics because for some reason those specifics matter to some people, and it's been personally beneficial to have more than an awareness of such legalese. Makes true believers feel I am living up to their "one of us!" means testing.

I accept they matter to others but reject such exists as anything but contemporary ethno-objects. Similar to how I acknowledge Christianity exists but am not a disciple of the dogma; I have to accept others believe but do not have to pledge allegiance myself.

Aside from engineering and healthcare, machine operation, with real safety implications, everything else is just parroting and social role-play.

Being a VHS copy-paste of generational semantics isn't the flex you want it to be. Patronize harder though.

Haha, this is excellent satire. Well done.

It's a benefit corp

Does that label have much impact on anything? Does it paralyze human nature?

They've been very clear about their mission, they're doing more than anyone else when it comes to it, and if you've ever interviewed with them you'd know how critical it is to them.

But I guess it's easier to make a glib comment than look these things up.

Amazing people lap this up.

Anthropic has one of the best moats of any business that's been created in the last 50 years.

Numerous companies have tried and failed competing with SoTA foundational models. If Anthropic had no moat, Apple and Meta wouldn't be paying them billions for coding asistance.

Meta, Amazon, Apple, and Nvidia would all have SoTA competitors to Claude. They all tried and have not produced a competitor.

Instead you have three companies that stand alone making billions from foundational models.

The open models are not far behind. Is it really a "moat" when it's so short lived and you need a brand new moat after six months? That's just ordinary competition.

They are far behind. Go check re-swe bench to see the overfitting measured

Or just try to use them. They don’t generalize as well.

They are benchmaxxed.

They're the least incompetent in the space.

Big companies are handcuffed by Innovators Dilemna etc.