Visually,

Red ones = Category leaders = NOT strictly-dominated

— Having some tradeoff that makes it superior in at least one way, versus all the others

Grey ones = Category losers = Strictly-dominated

— Being best at nothing derivative copycats. These you can expect to fade to zero, unless there’s somewhere that they excell that you missed.

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

13x $Butthole

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

BRUV

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

GM

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Zuckerberg:

"We're going to get rid of fact-checkers and replace them with community notes, similar to X."

"Fact-checkers have just been too politically biased and have destroyed more trust than they've created."

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Meet Web4

AI agents carrying out tasks on behalf of users, interacting with AI agents of other users, to negotiate and perform tasks, with crypto being used to pay for everything

For once, not bad

Some of us been calling for exactly this for years

Web4

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

AI Agent wave began taking off nearly 2 years ago

Then suddenly AI agents stalled hard

Stopped dead

Why? Not a real mystery

Viable AI agents needed just one more state-of-the-art upgrade to be viable

…but then, OPENAI JUST STOPPED MAKING ANY NEW SOTA MODELS, AT ALL, FOR 2 WHOLE YEARS

STOPPED. DEAD.

Until finally, just as competitors began catching up to GPT-4, OpenAI finally resumed work on the next batch of SOTA models,

e.g. the o1 line

With it, the AI Agent boom has finally resumed, as the github activity shows

And with that, the crypto “web4” boom has begun as well

OpenAI’s 2 year pause killed the boom for years, but now it has resumed

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Trudeau's mom

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

$BTC going down on that $100k line like Trudeau’s mom at a commie convention

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

$BTC back below $100k

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Yup

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

Remembering theorizing, back in September,

why OpenAI might have done the surprise move of pausing creation of any new foundation models, for 2 whole years

= Just to kill off 95% of the potentially-competing AI startups that had sprung up

^^ As can be seen from the 2022 tweets, many startups at the time had been making just that kind of bet, that OpenAI would keep continuously making new state-of-the-art foundation models, just as it had with GPT-1, GPT-2, GPT-3, GPT-3.5, GPT-4

But no, suddenly OpenAI just stopped dead, for 2 whole years, right when many startups needed just 1 more advancement the most, to fix some very fixable problems for agents that GPT-4 had

Intentional or not, OpenAI’s surprise 2 year pause certainly did have that effect of wiping out tons of AI startups, who’d bet that OpenAI wouldn’t pause

Fortunately, with OpenAI finally releasing the o1 models, the AI agent boom is now back

— and for good reason, AI agent boom is going to be massive

Already technically possible for these AIs to do FAR more than people realize, far more of actually economically valuable work

Likely requiring a bunch of AI inferences per each task, unlike today’s typical AI tools that just do 1 AI step per command — but in any case, now entirely technically possible

Get ready

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶

It’s over 3rd world bros

🄳🄾🄾🄼🄿🄾🅂🅃🄸🄽🄶