@levie
This wasn’t an issue specific to crypto or “fake assets”. Gambling with customer funds is illegal even if it were non-crypto, unless the customer agreements specifically allowed rehypothecation, and even then, you can’t just loan it all out to a non-arms-length sister company
@Dennis_Porter_
Huge difference. Banks are allowed to loan out customer funds, and it’s explicitly permitted in their customer agreements. Everyone knows banks make loans with customer funds. FTX committed fraud by loaning out customer funds. Their policies explicitly said they wouldn’t do this.
@yoheinakajima
We need to teach the robots how to love ❤️
Which, don’t get me wrong, will be no easy task, but seems pretty important if we want AGI to be friendly.
Thanks for the shout out!
@eigenrobot
Just wait until you find out that she’s not even dating him anymore — she’s scoring you free yardwork in exchange for setting him up with her friends. You find out a year later, and don’t know whether to be upset or impressed. At least you got free yardwork
@limegpt
@anothercohen
@rob_lh
People wouldn’t want to stay on a land resort with rooms that small, food that mid, etc. But on a cruise it’s okay because nobody expects land-quality on a boat
1/ I invested in Superbloom in August 2017. I was promised a refund in April 2018, which I never received. The following is from a Facebook thread with
@SuperbloomHQ
’s CEO Emmie Chang, after she said Superbloom was shutting down.
@fantasysalary
@billybinion
@jbarro
From GPT:
Certainly, here are ten more ideas for an ironic slogan:
1. "Who needs safety when you have character? Welcome to our 'not-so-safe' community."
2. "Embrace the chaos - Join us in our crime-ridden community."
3. "The only thing scary about our neighborhood is the crime
@Aella_Girl
I wouldn’t expect the deprived partner (regardless of gender) to just suck it up. To resolve, they either need to work towards an agreeable compromise together, agree on an alternative arrangement, or break up as a last resort. Cheating in this situation is still unethical
@pitdesi
@elonmusk
@sama
It kind of makes sense he has no equity. It would seem like massive self-dealing to have his non-profit give himself equity in the non-profit’s for-profit subsidiary
@yoheinakajima
To expand on this, not only do we need to train AI to have empathy prior to AGI, but it specifically needs to be *cross-species empathy*. Empathy alone isn’t enough if it’s not cross-species. Look at how humans treat animals. Strong cross-species empathy is key to friendly AGI
@potdoer
@cafreiman
No. Capitalism is people self-organizing their own groups, and some of them don’t want to take any risk or admin overhead in creating and managing the group, so would prefer to join an existing group. And if they don’t like their group, they can switch or even start their own
@brucefenton
I’d take Vivek over most candidates on either side now, even though I definitely disagree with him on many things. He is sane and rational, has charisma, communicates well, isn’t afraid to speak about controversial issues, and doesn’t seem corrupt
@RokoMijic
Centrally planned economies can’t take into account preferences and subjective value, which make up a significant portion of pricing and resource allocation
@Simeon_Cps
I even experienced this when explaining it to my wife. It took showing her a podcast with Eliezer as a guest for her to fully understand and accept what’s happening
@TylerAlterman
I have trouble knowing if any supplement actually works. It’s too subjective. I take the ones that have good evidence even if I can’t tell the difference. I think theanine has good evidence, just not as much as others?
Regardless of whether AGI should or shouldn’t exist, if people are trying to make AGI exist, we need to figure out how to teach it empathy. Ideally with empathy internal to itself, as part of its “brain”, not RLHF-trained after-the-fact. Thanks for the shoutout,
@yoheinakajima
!
Late night thought:
If LLMs or AI trained on our words are a true threat, and if the most powerful models will be trained on the latest data…
Then maybe spreading love and empathy is the best AI safety method. Every time we say something positive, we make it likely for future
🚨 Paul Christiano — who AGI accelerationists often reference when countering AGI doom arguments — thinks the probability of AGI doom is 50%. 🚨
This should be convincing enough for anyone who thinks AGI isn’t too existentially risky.
Meet the guy who invented
@OpenAI
's RLHF technique ().
His name is Dr.
@PaulFChristiano
, PhD Comp Sci from UC Berkeley.
Paul says there's a 50% chance of doom once AI systems are human-level intelligent.
So do we keep training smarter AI systems or no?🤔
@cafreiman
The enforcement of private property under the NAP isn’t a violation of the NAP. It’s the stealing or trespassing which is the violation, and the prevention of stealing or trespassing is justified — the only question is, how much force can be used? The minimum necessary to protect
@steinerwriter
@Austen
It depends if the private university is non-profit or for-profit. If it’s non-profit, there are no shareholders or company equity. Open AI was non-profit and now has a for-profit subsidiary with equity
A new survey just dropped and *wow* are Americans worried about AI:
- 82% of now favor slowing down AI vs 8% who support speeding it up
- 62% are primarily worried about AI vs 21% primarily excited
- 56% support a federal agency regulating AI vs 14% opposed
- 82% don’t trust
What will be valuable in a world where scarcity is a thing of the past? I think crypto and real estate will be the only truly scarce assets in an AI-enabled future. A thread 🧵 1/22
Met up for drinks tonight with Twitter friend and new in-person friend, Brendon
@PinkoPatriot
. Good times talking about AI, crypto, UBI, RCV, and more.
Any other fellow Twitter friends in Redondo Beach or Torrance? Let us know!
@david_perell
I used to do this at the Hyatt in SF. Such a great lobby. Used to often eat lunch there too, and then just stay at the table for hours working (and obv I’d tip well)
@visakanv
Broccoli, cauliflower, cabbage, Brussels sprouts, kale, collard greens, gai lan, and kohlrabi, aren’t just in the same family of plants (Brassica) — they’re actually the *same species* of plants (Brassica oleracea). Yes. The same species. All of them. 🤯
@SD_Better84
@wesyang
Are you claiming that Hamas murdering civilians intentionally, as the primary goal and not collateral damage, is not an atrocity?
@RealKidPoker
Instead of losing all your chips if you fall below 20 BB, what if you are instead required to go all-in on the next hand? Seems like that would solve the problem without having as much strategy needing to change
@AutismCapital
What about the EAs who would never dream of using it to wiggle out of moral responsibility? I think most EAs wouldn’t scam others or use EA as a moral cover
@balajis
Only $50mm was given to Democrats. How much was given to other politicians? And how much to nonprofits? I don’t think it was billions. Whatever the amount, it’s a small % of the money that was lost. It wasn’t even remotely a transfer of customer funds to these recipients
@michaelcurzi
Re: letting the baby cry at night and sleeping through it, health reasons aside, it’s too loud and constant to sleep through. Especially since most infants sleep in their parents’ room. To maximize sleep, parents need to feed/change/comfort the baby and get it back to sleep ASAP
@MikhailBatin
@bryan_johnson
I hope he releases more info, does studies, etc, but I don’t think he’s a scammer. He is legitimately doing an n=1 longevity study on himself using a variety of metrics, and open sourcing it. People are interested in this and there is demand for products that emulate his
@Bitbello
@AlyseKilleen
The fact that FTX invested in DeFi isn’t a failure of DeFi. The wrapped tokens are a good point though. Wrapped tokens are subject to centralized failures if they’re with a custodian and not a smart contract
@RichardHanania
Why does any of that matter in this instance? He was bothering people and yelling, yes, but this was not self-defense.
He was choked to death for the crime of public disturbance.
His other crimes in the past are irrelevant to this crime committed against him.
@NathanLands
Trying to learn and understand as much as I can. Trying to be stoic about the risk and instead be solution-oriented. Participating in the discussion. Not starting anything new long-term professionally yet, but still living my personal life the same as before. How about you?
@realDonaldTrump
While Dr. Fauci was calling for people to wear masks and stay home, you were brazenly not wearing your mask or encouraging others to, and calling for the “liberation” of all the states under lockdown.
@AISafetyMemes
Autonomous killing drones is obviously a catastrophic risk just waiting to happen. It’s almost like they’re trying to build Dystopia on purpose
I’m joining this funding circle.
I want to help increase the timelines to AGI, since I think the timelines are too short, and shorter timelines are too dangerous.
Is anybody in my network interested in joining as well?
I'm starting a funding circle for slowing down / pausing / stopping AGI development. Looking for people who want to commit $100k(+) funding this year to it. Please share with anyone you think might be interested in joining.
@eigenrobot
Have you never had perfectly prepared cauliflower or Brussels sprouts? Sautéed in olive oil with garlic and onions, served with pine nuts or almonds, or something like that. Vegetables can be amazing
This is wild. Even though these AIs are basically playing a video game, the fact that they can self-organize, have memories, figure out their own actions, etc, all with open-ended personas with no fixed goals… now we definitely know AI can be agentic
This is quite the paper!
It gave 25 AI agents motivations & memory, and put them in a simulated town.
Not only did they engage in complex behavior (including throwing a Valentine’s Day party) but the actions were rated more human than humans roleplaying.
@potdoer
@cafreiman
Countless small businesses, like mom-and-pops, freelancers, consultants, startup founders, auto repair shops, doctor’s offices, boutique attorney, and even lemonade stand kids, etc, beg to differ
@Austen
Still better than welfare, unemployment, and all the other entitlements, which are inefficient, have massive overhead, create a welfare trap, can be gamed, etc
@michaelcurzi
A great way for new parents to maximize sleep is to get the Snoo, a smart bassinet that plays white noise and rocks the baby to sleep. It adds about a half hour of extra sleep for parents per night by reducing by about one wake-up per night. I recommend it to all new parents
“Wtf, the AIs suddenly got...weirdly aligned?”
“Oh shit.”
“What??”
“Imagine you work for a paranoid autocrat. You decide to to coup him - wouldn’t you ease his fears by acting aligned?
So, AIs suddenly appearing aligned should set off alarm bells - what if they’re about to
I’ve explained why I think crypto and real estate will be the only truly scarce assets in an AI-enabled future.
If you enjoyed this thread and found it valuable, follow me for more threads about AI, crypto, and my vision of the future. And please retweet the first post!👇
What will be valuable in a world where scarcity is a thing of the past? I think crypto and real estate will be the only truly scarce assets in an AI-enabled future. A thread 🧵 1/22
@PatchBandana
@limegpt
@anothercohen
@rob_lh
Cruises optimize public spaces pretty well, but not private spaces. Standard guest rooms are so small that there is barely room for a bed, and the bathrooms are so small that the shower and the toilet share the same space.
Making guest rooms on a cruise bigger would
@michaelcurzi
"This may seem expensive but it is an extremely cheap way to purchase everyone having a great time and they'll love you for it"
Words of wisdom!
@cafreiman
Preventing a non-ticketed person from boarding a train isn’t coercion, it’s prevention of trespassing. Blocking a pedestrian on the street isn’t morally or practically the same as blocking a non-ticketed person from boarding the train
the fight for the future is a struggle between technological-driven growth and managed decline.
the growth path has inherent risks and challenges along with its fantastic upside, and the decline path has guaranteed long-term disaster.
stasis is a myth.
I hooked up an iOS shortcut to GPT-4 and GitHub.
Here’s how it works:
- identifies the repo
- identifies the target files
- writes the code for me
- creates a branch
- commits changes
Now I can code right on my Apple Watch with nothing but my voice!
@RokoMijic
@r1p0st3_crypt0
@Freakoutery
It’s an interesting idea, but the loser would only surrender if they trusted the simulation, which may require them being stakeholders in creating and running the simulation. You’d need to effectively get warring parties to agree to use the simulation as binding arbitration