Every AI company knows they're building the wrong thing.
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan It's being going on like that for years... I wonder when everything will fall apart because of very very rotten and bad foundation...
I like your ending though. It gives me some hope...
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan One challenge I see is that LLMs are really good for copying things because an existing product can essentially be a spec, which makes me think small companies are going to get with a worse version of the app store cloning problem.
This seems especially bad for tools which the company they're renting services from might want: Mac developers talk about getting Sherlocked but that was limited by developer time in a way that is very different now.
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan It's like a space race to get to a planet that doesn't exist.
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan
You are right
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan It is not artificial intelligence it is Computerized Lying.
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan
How can we protect ourselves from being the victims of the wrong thing that the AI companies are creating for their own benefit? -
@Daojoan Thank you for this essay. You have put your finger on the "herd behaviour" of SV tech.
Joel Spolsky did NOT say everything you are now saying twenty years ago, not even close. But his "Fire and Motion" essay from 2002 is a complementary read: https://www.joelonsoftware.com/2002/01/06/fire-and-motion/
@raganwald @Daojoan
Moloch… -
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan I agree 100%. For the better part of a year now, it feels like the expectation is that they'll just *happen* upon an ultimate AI product. The target doesn't exist, but they expect a bullseye.
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan
It really is reminiscent of the tulip bubble, in that you could genuinely go out of business if you didn't participate. It's a giant game of financial chicken. -
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan 'AI' seems uniquely bad because of the very high cost of entry. VCs are far more cowardly herd animal and much less adventurous than advertised across the board; but that matters less when the cost of entry is close to zero and the burn rate is a handful of developer salaries; while it is the single concern when you need to set a billion dollars on fire just to sit at the table.
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan CLI then ... oh ... claude ... + api ... oh chatgpt has it...
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan
I've heard a million times that AIs write better code than people, so these AI companies should be using an / their AI to write their AI. It should speed things up and save them money. -
" Find a niche that's too small for the Moloch-captured companies to notice and go so deep into it that by the time they arrive, your product is years ahead in the dimension that matters to those users."
Yes.
Then you run into what I call "The Oracle Problem". Eventually an Ellison (or Jassey or Pichai or Zuckerberg or whomever) will notice your product is profitable and do one of three things:
Buy you out - whether willingly or hostile
Sue the shit out of you for whatever bullshit reason they can come up with and keep fighting you until you run out of money
Drop a bunch of money to make a shoddy competitor but push it to your customers for half the cost or as part of their existing bundle or whatever until you go under
In the long run, the rich people win
-
Every AI company knows they're building the wrong thing.
And they can't stop.
Because the competitive structure specifically punishes anyone who slows down long enough to build the right thing.
The race is real. The destination is fake.
https://www.joanwestenberg.com/everyone-in-ai-is-building-the-wrong-thing-for-the-same-reason/
@Daojoan i mean... there's a lot of truth to this. But then also, at the frontier lab level... Yann LeCun? Ilya Sutskever? And do we really believe that Google Deepmind are solely focused on LLMs and the chat interface for them?
-
J jwcph@helvede.net shared this topic
T tokeriis@helvede.net shared this topic
