As a retail investor, you have an obligation to never buy Anthropic stock. They held out in private markets, shopped their IPO all over the world before finally trying to dump it on you at >$1T valuation.
Do no reward them by buying their shares. Do not be their exit liquidity.
I can't believe the banks are down and I can't access my money.
Can someone please invent a global, 24/7 network, with a redundant ledger that enables self-custody of my funds and is resistant to state-level attacks?
NEW: @telegram CEO @durov ANNOUNCES ANY TELEGRAM MESSAGE CAN NOW BECOME AN APP
DOWNLOADS INSTANTLY (JUST 4 KB)
INTERACTIVE AND NATIVELY RENDERED
PAYMENTS AND VIRAL MECHANICS BUILT IN
Ceding Ukrainian land for peace is just abstract words. It’s not only about reality, even more about philosophy. Putin was talking about Donbas, but then he decided to occupy Kyiv, our capital. His words were about Donbas, but the reality was about Kyiv. He kept saying he was defending Russian-speaking people and then tortured hundreds and thousands to death in Bucha and other places. So that wasn’t about Russian-speaking people.
This is why when Americans asked, “Can you exchange Donbas for security?” I said right there, “Can you tell me what it means? What security? NATO?”
“Not NATO, but similar to NATO.”
“Similar? But you don't want the answer about Donbas to be ‘similar.’ You want a clear answer about Donbas. This is your goal. And this is Putin’s goal, for today. What will you do if we have security guarantees but not NATO?”
They took our nuclear weapons. I'm not a fan of nuclear weapons. I think the world would be a better place without them. But the Russians took our nuclear weapons, didn’t they? They did. Then they stopped our integration into Europe, then prevented us from being in NATO. And in the end they occupied us. This is why it’s an abstract question. Nobody trusts Putin to stop if we leave Donbas. He will pause. Then he will come back in one year, two years, or five years. And who will be guilty?
From an interview with DR Nyheder. (2/3)
“I see no reason for this trend to stop as long as the supply of memes and launchpads continues to expand.”
anyone could see this from a mile away but the obscene amount of money poured into the launchpad products and eco, as it prints fees for parties involved, is whats killing it. It will continue to be viable as long as odds of “winning” are better than buying a lottery ticket.
Cant wait for this segment of the “asset” class if you could call it that, to finally fully die.
Interesting to reflect on the latest loop of the euthanasia coaster in this recent run.
Top memes couldn’t even hit $500m mc, a “runner” was like 30m mc.
Sad state of affairs comparing to the glory days.
I see no reason for this trend to stop as long as the supply of memes and launchpads continues to expand.
Yeah it’s fun to punt a few bucks at these things and am all for it myself, it’s fun.
But don’t let the industrial trench farms psyop you into thinking any of this is real or durable
a lot of tech people and especially VCs are very disconnected from the average consumer even the average business
im surprised that meta went to make muse just following the trend - they should have all the data in the world to tell them how people shop and that there’s only a small percentage of the population that would benefit from having an AI assistant, and that % of the population already can hire a human
obviously the technology is cool but it’s the same as with the meta verse and the VR bet, the question is going to be: what is meta going to do with all the compute they are building?
We benchmarked frontier AI models on reconstructing 3D characters from 2D references in Blender.
Our findings reveal limitations in spatial reasoning and geometric reconstruction, alongside blind spots in conventional 3D evaluation metrics.
Glad to be able to collaborate with @tpavankalyan23 and team at TensorTest on this.
https://t.co/JwzkReejbn
Automatic scores catch wrong proportions. They miss what an artist would actually fix: the expression, how hair meets a head, details the artist never drew.
Next: can a model fix its build from an artist's notes?
Study + comparisons: https://t.co/Dm8J6H4J92 Thanks @next_roy
How well do frontier AI models actually understand 3D space? We collaborated with TensorTest to benchmark 3 models on character reconstruction.
Some interesting gaps in spatial reasoning, geometric accuracy, and how we evaluate both.
https://t.co/gPQmpWVsNW
Automatic scores catch wrong proportions. They miss what an artist would actually fix: the expression, how hair meets a head, details the artist never drew.
Next: can a model fix its build from an artist's notes?
Study + comparisons: https://t.co/Dm8J6H4J92 Thanks @next_roy
The hardest character was a squat shark in a jacket with a huge collar. No model got past 42%.
Two tied at 42% for different reasons: one gave him a thick, even row of teeth, the other a thin slit. An artist would send them back with different notes.