Husband, OCM holder, poker player, curious about AI safety, pause AI, lawyer, living in Montenegro https://teapaw.org/t.co/2JbkdNmxt2 https://teapaw.org/t.co/GsWUAS7YQM
Monkeyverse
Joined November 2021
- Tweets6.7K
- Following86
- Followers856
- Likes21.3K
Pinned Tweet
Hello 👋 from Montenegro 🇲🇪 my friend, the Adriatic sea 🌊 has become colder, but I am still swimming 🏊 every day. What’s up in your lands?
Margot retweeted
Adapting to cold with contrast showers 🤝 Started 2 weeks ago with 30 seconds🥶, now can tolerate up to 90 🤘
Margot retweeted
6 months. Nick Bostrom is possibly the most unknown expert on superintelligence. In fact, he popularized the risk and the term in his 2014 book, Superintelligence.
As an Oxford professor, he has always been understated p-doom percentages and timelines, but he says now it could arrive even in 6 months.
That's something that should put us all in emergency mode and act on those timelines.
Nick Bostrom says Superintelligence could arrive in 6 months, and that it poses an existential risk 😱
"It's really the last invention we will ever need to make... but there are also significant risks including existential risks."
"I think we have to take even short timelines seriously, like, you know, could happen in 6 months, a year or two."
Margot retweeted
Depends on the table: it’s easy to pay the rake versus -200bb/100, it’s harder versus -40bb/100, and it’s much, much harder versus about zero regs. Everyone deserves a different individual mood, style, and share of concentration.
Also, some interesting lines are more interesting to test versus recreational players even when you have a D-game mood. What do you think? How do you properly bring up your D-game when you want to watch a movie or YouTube video but still need to play?
For example, today I wanted to rewatch Monty Python and the Holy Grail because I wanted to better understand King Arthur for Halloween 🎃. After a short good 15 minutes of play, I lowered a few limits and played only one or two HU tables that I usually play.
But anyway, I don’t really enjoy poker and would be happy to do something else.
It’s so hard to focus on it seriously with total ignorance of other life
Sadly biggest bitcoin vault is not safe,
It’s not first time when Binance was hacked, they usually stop answer to police when it’s become obvious than they leaked security responsibilities of themselves,
CZ answer that Binance would pay to consumers also usually do not make Binance to pay even when he was CEO,
And main worry is that Binance is biggest BTC holder and so big that it couldn’t be completely safe
The Binance hack so far:
- attacker deploys a sweeper contract, gas paid through Tornado Cash
- attacker’s address starts receiving funds on Base, then Optimism
- 18,270 BTC leaves a Binance hot wallet
- 612.7M USDT and 46M TRX are taken on Tron
- Binance 14 gets pointed at the sweeper via EIP-7702
- one call sweeps 167K ETH and 1.6B USDT from it, $2.54B
- Binance 16, Arbitrum and Avalanche wallets send more
- 823M USDC is pulled from a Binance smart account
- Tether starts freezing, 1.57B USDT in total
- Binance 14 gets swept a second time
- Circle freezes 823M USDC
- CZ: “It does look like a hack on Binance”
- attacker moves unfrozen tokens and starts selling UNI, PEPE and RLUSD
- third sweep takes 200M USDT, which lands frozen
- Binance halts withdrawals until the security issue is resolved
- Ripple freezes 40.5M RLUSD
- Tether blacklists Binance 14 itself
- WLFI and BitGo freeze 27.8M USD1 and 48.5M WLFI
~$5.8B taken. ~$2.7B frozen. ~397K $ETH and 18,270 $BTC still with the attacker.
Binance 14 is still pointed at the attacker’s sweeper.
Margot retweeted
I didn’t really buy the 50/50 predictions — they usually feel more like rhetorical positions than actual forecasts.
But I agree with the broader point: if no one builds advanced AI, we still die. Even with a relatively modest 1% annual risk of catastrophic damage (nuclear war, climate, bio, cosmic, etc.), the chance of surviving 100 years in anything like a decent state is 0.99^{100} \approx 36\%. That’s already a lot worse than most people intuitively want.
We’re near-certain to die as individuals anyway, and the collective risk over a century is substantial. Progress in any direction carries risk. So at some point we have to accept that and keep going.
I also don’t see much benefit for humanity in pure worrying. There are real problems with AI systems, and not enough serious work on them. At the same time, collective government agreements could meaningfully raise the bar: countries could formally treat the creation of any product with clear extinction potential as a crime against humanity.
Right now the Rome Statute’s definition (Article 7) doesn’t cover this — the list is closed (murder, extermination, enslavement, etc.) and doesn’t include developing technologies capable of destroying humanity. The Statute itself isn’t universal either: roughly 125 states are parties, major powers never joined, and several countries have withdrawn or are in the process of leaving. It would also help to have proper procedural innovations for arrest and prosecution of such actors, ideally with every country signed onto the same framework.
Margot retweeted
While I sympathize much more with Elijah Yudkowsky and Netzaris's preference for humans among existing and future sentient beings, I consider myself a specist. I do not consider myself a long-termist, whereby I believe that current humans deserve much more attention, care, and preservation than future possible humans.
I respect all their positions, but what comes to mind is what Nick Bostrom said the other day and wrote. If the book by Eliezer and Nate was called "If Anyone Builds It, Everyone Dies" (in reference to superintelligence), Nick Bostrom said that there's a much truer statement that is the exact contrary: "If no one builds it, everyone dies." I can't avoid recognizing the fact that he is absolutely true. His statement is true 100%, or, actually, it is very likely true.
It's not necessary, because there are ways to achieve even long-term philosophy, escape philosophy, and velocity, even without AI. I want to go to AI because, yes, we may achieve that, but not other goodies, which may actually be a precondition for long-term longevity to be a good thing.
Anyway, I thought it was interesting. What he said is possibly true. Actually, they're both partially true. The first one is partially true because if anyone builds it, everyone dies. We don't know if everyone dies. Maybe 50% of everyone dies, and 50% of everyone doesn't die. There's a good chance of scenarios where nearly everyone dies, which is quite different from, number-wise, it being pretty much the same thing.
It could provide for such a way that, at some point, as this superintelligence has developed, it could then give out to this nucleus of humanity (which would also include the cultural background and so on) the ability to extend itself again, on and on in a perfected manner, so a manner that possibly preserves our authenticity but helps us fix our educational system and so on, and cultural background, and have them redevelop around a solved race. Raise for resources, for survival, and for wisdom.
1/10 Just translated one of the most legendary cash game videos ever made — plus the full translation of core blueprint
By Nikita Bodyakovsky (Mikita Badziakouski), one of the most successful poker players of the modern era. He’s currently top-5 on the All-Time Money List with over $68 million in live tournament earnings, a WSOP bracelet winner, multiple Triton Poker champion (including Main Events for $5.2M and $2.5M), and the undisputed #1 Belarusian player of all time. Many consider him one of the founding fathers of the Belarusian poker scene and the core figure behind the “Red Army” movement that an entire generation of players grew up on. In this first 30-minute part you get a complete, no-nonsense guide on how to exploit recreational players in home games — one of the main reasons Belarusians became one of the very few nationalities that consistently crushed the games for years.
Now available in English for the first time.
(you could find link to video on 10/10)
9/10 In general against fish we always choose sizes that are profitable for our hand + that support our plan for the next street. For example:
if we plan to under-bluff the river against this fish, we bet denser on the turn
if we do not plan to value-bet the river, we again bet denser on the turn
if we have vulnerable value, we again bet denser
if we have a hand protected from all sides (pair + draw), we induce with small sizes
if we have a plan to keep value-betting the river, but the hand is not nuts (strong top pair), we bet smaller instead
10/10 Part 1 of the video: youtu.be/8NU-jf2m3ww Please repost 🔃.
You know who needs this.
You shall send it to them. 😀