cypherpunk.kas retweeted
we are truly in the end times
EXCLUSIVE: I'm 14, and I just interviewed Sam Bankman-Fried.
We talked about:
– Whether FTX was worth it
– His p(doom)
– Why he only put $500M into Anthropic
– Whether he’d flip a 51/49 coin for his own freedom
– How FTX almost became a political prediction market
– Why he thinks we should “get rid of the feds”
– What he’d do if he were my age
The Supercycle: The Sam Bankman-Fried Interview is out now.
00:00 Highlights and introduction
01:52 Life in prison and the justice system
11:09 Anthropic and the future of AI
20:00 Investing in Anthropic and running FTX Ventures
29:25 Effective altruism and the FTX Future Fund
35:33 AI safety and the race to superintelligence
40:32 AI timelines, extinction risk, and mind uploading
50:43 Prediction markets and the future of finance
57:22 Books, learning, and entertainment
1:00:22 The coin flip that could destroy the world
1:06:52 Slowing AI and the limits of expected value
1:14:38 What actually helps AI safety?
1:17:40 Was FTX worth it?
1:19:28 Advice for young people in the AI era
cypherpunk.kas retweeted
holy shit.
16GB of RAM.
that's all you need now to run an AI that codes like sonnet 4.5 at home.
Some of you thought that Inkling 2 Small was too big at 594B parameters. We heard you: Inkling 2 Mini is a looped transformer with just 26B parameters, small enough to run on your laptop, powerful enough for agentic work with a score of 20.1 on AA Index.
thinkingmachines.ai/news/ink…
cypherpunk.kas retweeted
gpu's can now make enriched uranium!
The Riemann Hypothesis has fallen. RH was the singular math result that had been named, in advance, as the obvious coal mine canary for dangerously advanced AI, over the last three decades. The methods used should generalize to "make AI very good at anything you can definitely test", and unfortunately that category includes "make AI better at making AI".
This is like a private company announcing they have produced a million tons of 15% enriched uranium. It's not 90% enriched weapons-grade uranium, so you could proclaim the company's output is not immediately deadly right now. The concern is that the process line may generalize, to an extinction-level threat, within the unpredictably immediate future. Forecasters still had Riemann at 2035 one week earlier.
This is distinct from earlier events in that earlier you could reasonably ask if AI was selectively picking off famous math problems that turned out to be easier for AI. The Riemann Hypothesis was the single challenge most often named in advance as a warning sign.
cypherpunk.kas retweeted
We got GTA 6 before we got $TIBBIR exit from stealth.
cypherpunk.kas retweeted
JUST IN: Riemann Hypothesis has been solved.
alpo.ge/remark.pdf
cypherpunk.kas retweeted
Wrapped tefillin the legend, @mickymalka, the greatest FinTech investor of all time and founder of @RibbitCapital, where he manages over $20 billion.
Micky’s entrepreneurial journey is remarkable.
His passion for investing started at just 13 years old.
Growing up in Venezuela, shortly after his Bar Mitzvah, he used his Bar Mitzvah money, borrowed the rest from his grandfather (who even charged him interest!), and bought his very first share of Berkshire Hathaway.
He later went on to co-found his first financial services company as a teenager.
He then Founded Patagon.com, which Banco Santander acquired in a deal valued at approximately $750 million in 2000, when that was a lot money.
In 2012, he founded Ribbit Capital, which went on to make early bets on the biggest names in fintech:
Coinbase.
Robinhood.
Nubank.
Credit Karma.
Brex.
Revolut.
But beyond the extraordinary business career, what stood out to me was our connection over Jewish identity and pride.
It’s inspiring to meet Jewish leaders who have helped shape the global business world while remaining deeply connected to their Jewish roots.
cypherpunk.kas retweeted
Quantum resistant addresses are fully supported within @KaChat_ apps!
Anytime you see a green shield on an address it means its quantum resistant. Be mindful that you lose resistance whenever $KAS leaves an address for the first time.
Fun fact: quantum resistant addresses start with kaspa:p while non-resistant addresses start with kaspa:q
cypherpunk.kas retweeted
JUST IN: Over 400 Anthropic employees are now worth more than $100,000,000 each following the $ANT IPO.
Anthropic's seven co-founders are now worth over $20 billion each.
The average Anthropic employee is worth over $40 million.
cypherpunk.kas retweeted
Anthropic's IPO price is overvalued. It currently stands at above $1 trillion.
It's overvalued because over time, most inferencing will be done by smaller models that will fit inside smaller compute platforms.
AI will end up looking like personal computers and not like central cloud servers.
In that case, the value capture of any one single AI company is going to be much smaller: it will be commoditized.
This is not investment advice, DYOR.
cypherpunk.kas retweeted
JUST IN: 🇺🇸🇷🇺🇨🇳 Trilateral meeting between President Trump, Xi Jinping, and Putin officially begins.
cypherpunk.kas retweeted
Replying to @Justin_Bons
SOL’s fast finality requires a supermajority of the stake to sign off on the state. If SOL were properly decentralized (in its most important security axis: number of independent stakers), aggregating that agreement fast enough would be the processing bottleneck. That’s a tension which seems unavoidable in the PoS model.
(teapaw.org/michaelsuttonil/status…)
The case for the uniqueness of fast pow
tl;dr
Finality has two moving parts: (i) fast inclusion (= high bps, how quickly a tx gets into a block), and (ii) fast confirmations (= how quickly that tx becomes irreversible). Any system with rapid block production can achieve the first. The second is where the tension shows: in pos, fast confirmations press directly against decentralization. In fast pow, the two properties are decoupled.
prologue
A few weeks ago I came across Solana’s founder claiming: “Solana is the fastest monetary system in the world”. Since Kaspa already runs at a faster block rate, I was curious to check Solana’s finality times. That curiosity quickly pointed me to a deeper issue: not raw speed, but how speed interacts with decentralization.
——————
The tension is structural. In pos, finality means accumulating staked votes, and the more decentralized the stake distribution, the more time is required to reach finality. Here I’m not talking about hardware requirements or validator specs. The axis I’m discussing is centralization around the security mechanism itself: stake in pos vs. hardware in pow. To be secure, a block must be confirmed by a supermajority--typically >66.7% of the total economic stake. In a truly decentralized network, where n stakers with uniform share grows without bound, the time to coordinate this supermajority becomes a real bottleneck.
Pow works differently. It samples the hardware space without requiring the protocol to explicitly collect evidence from a majority of miners. Each block is itself a statistical proof that the finder out-competed the full network’s hash power. This process--and its timing--remains independent of how many individual miners participate.
Ethereum’s researchers understood this when moving to pos. Unlike Solana, which tolerates concentration to reach ~13-second finality, Ethereum’s designers could not accept that trade-off. Their solution was to introduce rotating committees. A rotating committee is a smaller subset of validators, randomly chosen from the full set, that votes on behalf of everyone else.
But this comes with a different security model, known in the literature as exposure to a BFT adaptive attacker. The committee is selected first and then votes. That “select-then-work” sequence is theoretically exposed to adaptive attackers, since members are known in advance. Pow, by contrast, is “work-then-select”: the winner is only revealed after the work is done. Think of it this way: in pos, you know who the referees are before the game starts, which gives an attacker time to pressure them. In pow, you only learn who won after the work is already done, which removes that attack surface. So n confirmations provide consistent confidence regardless of miner granularity, and the system stays secure even under adaptive targeting.
Beyond attack subtleties, the real issue is economic weight. When I send a billion-dollar transfer in a pos system, the question I care about is simple: how much stake is actually securing it? A committee vote provides strong statistical evidence, but only a true supermajority puts the full economic stake of the network behind my confirmation. In other words, a sampled committee may convince me that things are probably safe, but only the weight of the entire stake provides an overwhelming guarantee. And this is exactly where pow shines: each confirmation is not just a probability estimate, but a direct proof of work done against the full hash power of the network, no matter how many miners there are.
closing remark
I don’t claim to know every engineering detail of Ethereum or Solana. But I’m convinced the core principle holds.
I’ll state it simply: fast pow uniquely enables fast finality without forcing a compromise on decentralization.

































