All of Charles He's Comments + Replies

Hello, if you plan on making further interactions on the forum, you might consider it useful to add your role to your profile, this would make your interactions more clear to a wider audience.

The link is probably here: https://forum.effectivealtruism.org/profile/molly/edit (this only works for you obviously).

Thanks! Updated. 

Yes, I think we're agreed.

it's misleading to think of FTT price dropping as "the thing that was life or death for them." (Or maybe it was in a "proximate cause"  kind of way, but the real problem was the reliance on FTT in the first place.)

It's more like FTT was a quasi peg, they needed to keep it up at $>20. The fight that was life and death was keeping it that high with their resources.

Fair enough. WSJ story here saying that Alameda owed FTX $10B!

Yes, as you say, the FTT token wasn't worth anything even before the crash, the FTX/Alameda money to prop it up is what was operative, and is gone now. 

We haven't discussed anything that would contradict the overwhelming evidence that FTX has a gap of $4B or more. 

Low information threads seem undesirable if there are people who are less informed and had very high/trust in SBF, partially due to EA associations.

Are you saying you have the balance sheets for FTX? Can you link? 

Unfortunately, I think what you're saying is omitting liabilities, and that makes it very uninformative.

5
Greg_Colbourn
1y
The ones that were reported by Coindesk and others last week, for which their legitimacy wasn't disputed (although here Caroline claims there is more). If it was just a liquidity issue, surely you'd think they would've been able to fix it by now. It's quick and easy to sell crypto tokens, even OTC.

For Alameda, other "coins" were covered in the link in my first post, it's pretty clear that they aren't worth anything, even if there was no crisis. 

https://dirtybubblemedia.substack.com/p/is-alameda-research-insolvent

This is probably true of any "projects" on FTX that the entities control.

6
Lukas_Gloor
1y
Yeah but the same is true of FTT under the assumption that FTX/Alameda rely on FTT for emergency liquidity.  I guess if someone had faith that FTX/Alameda would never sell a lot of it at once, but instead slowly sell over many years while keeping the exchange operating with profits, then FTT could be worth something for buyers.  But on this model it doesn't make sense to use it as collateral, let alone in any relation to backup for customer funds.  (TBC, we're not 100% what happened, but if FTT was involved in securing customer funds, that's very dumb at best and quite possibly illegal, as discussed by Matt Levine.) You probably know all of this – I'm just commenting because IMO it's misleading to think of FTT price dropping as "the thing that was life or death for them." (Or maybe it was in a "proximate cause"  kind of way, but the real problem was the reliance on FTT in the first place.) 

No, not at all. This "sale" already happened. 

Right now, the tokens are "illiquid" in the same sense that "Charles He token" is illiquid, "depending on price".

In theory, FTT at >$20 would support the whole endeavor. Whatever happened was life and death and happened two days ago.

Whatever happened was life and death and happened two days ago.

I'd say the thing that was life and death for them wasn't so much the price of the token (that was only a trigger) but the bank run that came after the token situation hit the news. Even if the token had stayed at the same price temporarily, no one could seriously expect their stake to be worth "number of coins times price at the time" (or 50% of that, which one source reported they had "conservatively" marked it down to) given the low liquidity  / low historical sales volume of the token,... (read more)

7
Lukas_Gloor
1y
They probably have on their balance sheet other illiquid low-circulation coins with inflated market cap where they were early investors or even (partly) coin developers, so it's possible that the claim was technically true at the time Sam stated it.  Of course,  it's an annoying game to play when you can't assume that people communicate with an intent to convey all relevant information as clearly and comprehensively as possible, so if we have to go to these convoluted interpretations, so much has already been lost. 
7
Greg_Colbourn
1y
I interpreted it as him saying that, even with FTT 80% down (and similar for other holdings), they still have enough to cover  customer balances. And I'm saying that would only be true in theory as the price could still go down a lot further (and would if they liquidated all their holdings). According to the  balance sheets, they held something close to the entire marketcap of FTT as based on circulating coins even before the crash. It's unlikely they've sold even a tiny fraction of that.

This is very sad and terrible, but I think the relevant companies are bankrupt and rescue seems unlikely. Maybe if it's helpful in some way, the newspapers Reuters and NYT contain information, if you want to read it. 

 

Onlookers: Guys, it's a highschooler, I think there might be special duties  here to people of different ages and backgrounds.

2
JoyOptimizer
1y
While they are insolvent, FTX and SBF have not declared bankruptcy. In developing scenarios, information is unclear and from unknown sources. (Alameda's balance sheet may prove incomplete.)

For onlookers:

Basically the tweets promote the narrative that it is a short term liquidity crisis and there is a future for FTX.

As someone who strongly supports SBF, it’s perfectly clear that FTX, and Alameda is in the beyond dire position of insolvency as reported.  This tweet is performative and misleading but is the best narrative the CEO can do, this fact is also is expected and normal at the same time.

2
Greg_Colbourn
1y
"4) FTX International currently has a total market value of assets/collateral higher than client deposits (moves with prices!)." - I think this is in line with the mainstream narrative of the assets (e.g. FTT) being almost completely illiquid - i.e. no way they can sell enough without crashing price to ~zero. So kind of misleading to say it's liquidity when it's really (most likely) insolvency in practical terms.
6
David Mathers
1y
Why doesn't lying about this expose him to more legal risk? Feels like the sort of thing that ought to be illegal. Or if it does, why bother? EDIT: To be clear, I don't doubt he is lying. It's just if everyone knows it a lie, and it looks bad in court, why is this the done thing? 

What I think: I think that FTX was insolvent such that even if FTT price was steady, user funds were not fully backed.

Yes you are right, I disagree. I think this collapse happened because of the FTT "attack" (or honestly, huge vulnerability) and Alameda was forced to defend. Without this depletion, SBF or FTX could cover these funds in a routine sense and we wouldn't hear about this. 

That is, they literally bet the money on a speculative investment and lost it, and this caused a multibillion dollar financial hole. It is also possible that some or all

... (read more)

I'll escrow with Nuno at any time, you can reach out to me by PM or Nuno can reach out to me. 

Given that you think it's likely FTX "gambled" user funds I am really not sure we disagree on anything interesting to begin with :-[ 

I think this is true, but sort of productive to clarify this.

Maybe you think it's only 70% likely and I think it's a lot more than that? 

I'm >90% sure FTT was collateralized in a way that had "a major role" in FTX's collapse, but like, I'm not sure exactly what that means in a causal way, and how severe it is in a moral sense. Financial engineering is complicated and I don't know much about it.

Maybe we disagree on just how big FTX's financial hole is? Could we bet on "as of today, FTX liabilities - FTX liabilites >= 4bn"? I'd go positive on that one.

It's hard to say, 4B is the ask but 8B was mentioned as a figure too. They could mean the ability to deploy some latent funds in some complex way. Honestly, this doesn't seem that meaningful.

Dunno... Really can't tell what you believe. You commented that folks are being too negative yet seem to also think that FTX "gambled" user deposits, which sounds pretty negative to me (though we can disagree

... (read more)

What I think: I think that FTX was insolvent such that even if FTT price was steady, user funds were not fully backed. That is, they literally bet the money on a speculative investment and lost it, and this caused a multibillion dollar financial hole. It is also possible that some or all of the assets - liabilities deficit was caused by a hack that happened months ago that they did not report. 

As far as I can tell, you don't think this. Well, if you really don't think that, and it turns out you were wrong, then I'd like you to update. I think probabil... (read more)

Wait...(the experience of actually putting real money seemed appalling and sort of scary, so I was looking through your profile to see if I would wiggle out of it.) 

You're Agrippa! The guy with very short timelines, is Berkeley adjacent and knows that cool DxE person. 

No, I do care about you! I respect you quite a bit. I was wrong and I retract what I said before in at least a few comments, and I apologize for my behavior. I upvoted every comment in this chain of yours. Also, I'll be happy to take any negative repercussions.

 

For the bet, I'l... (read more)

2
Agrippa
1y
😳 That's nice of you, thanks. I'm actually not a guy though I don't take any offense to the assumption, given my username. Maybe Nuno would escrow for us.    I'm probably down for $500, would need to talk to my partner about going much higher anyway. If you are in the US we might not need escrow since suing eachother is an option, if we went >5k that would be worth it.  Re SBF vs FTX/Alameda paying: Yeah I meant SBF personally. I agree it's a big difference. Jan 1st is the date but I also don't know how fast this stuff ever goes and researching it sounds annoying.  Given that you think it's likely FTX "gambled" user funds I am really not sure we disagree on anything interesting to begin with :-[  Maybe you think it's only 70% likely and I think it's a lot more than that? 

I'll announce here that I'll take positive, will resolve affirmatively, 50/50 on both of those, under the protest that "gambling" actually means "any collateralization strategy that failed". 

I'll announce, for 50/50, I'll take positive, will result affirmatively on: 

  • By April 2023, will evidence come out that FTX gambled deposits rather than keeping it in reserves?", 

Under the protest that "gambling" actually means "any collateralization strategy involving FTT that failed".

For 50/50, I'll take negative, will not resolve affirmatively on:

  • "SBF
... (read more)
4
Agrippa
1y
Cool, what size bet? And, after we figure that out, any thoughts on an escrow? 
3
Agrippa
1y
:-(  I will have to insist on trusted escrow for any bets between us...

You still haven't given an operationalization and at this point, we both know why.

I'll take a bet when Nuno or someone with a real name comes along.

5
Agrippa
1y
We seem to have very different ideas of what "operationalization" means... How about "By April, will evidence come out that FTX gambled deposits rather than keeping it in reserves?" ? There's already a literal prediction market up on that one!  We could do "SBF found guilty of literally anything / pays a fine of over 1M for literally any reason, by 2024" ? If that's not operationalization I really have to give up here.  I do have a real name by the way!! BTW I am assuming you are willing to bet in the thousands. If not, I really don't consider that a bad thing, but lmk please!
8
Agrippa
1y
As an aside it is surprising to me that I seem at all to you like the type of person Sam might have been surrounded with. I don't think anyone remotely insider-y has ever even slightly felt that way about me. 
3
Agrippa
1y
OK whatevs, which side of 50/50 do you want? And by what date? (and for that matter what X? Fraud???) That said I really dunno why you don't like "FTX used user funds to make risky investments" or "FTX speculated using user funds" etc. Is there nobody we might mutually trust to neutrally trust to resolve such a thing? 

"FTX committed fraud that caused >1bn loss of user funds"

We, and really, every prediction person here, know that's not an adequate operationalization. 

You've been asked 4 times already to give an operationalization and number for me to take, which we both know you are fully able to do.

 

I suspect you are an FTX grantee. Since you are using EA money (well until you decide it's not convenient to call it this) do you mind sharing who you are or what your project is?

6
Agrippa
1y
I'm sorry but I really don't understand why you think it's not adequate. "Fraud" is quite well-defined, and "loss of user funds" is also quite well-defined.  I would offer odds on like, criminal prosecution results, but that will take such a long time to resolve that I don't think it makes an attractive bet. As you point out there are also jurisdictional questions. Is "SBF lied about the safety of user funds on FTX.com" better to you?  "FTX used user funds to make risky investments"?  "SBF mislead users about the backing of their FTX.com accounts"?
7
Agrippa
1y
The real world event would be "FTX committed fraud that caused >1bn loss of user funds". But if it's a bet somebody has to arbitrate the outcome, you know? I just picked EA forum users as an arbitator since like, that's the venue here. But if you have any other picks for arbitrator that would be fine. You can pick yourself but I'm not sure I'd agree to that bet. Likewise I assumed you wouldn't go for it if I picked me. And if it's the 2 of us well then we might tie. > but do you have like an an account on a prediction market Multiple  > Are you associated with the grantees on their prediction market projects? I am not sure that I understand this question. I have myself recieved a grant to work on a prediction market project. 

I don't really have much to say except pretty much copy and paste my reply, above and point out how wildly hostile this is. No, people are not expected to put up bets for their statements. You are perfectly able to dispute them using arguments. 

This is despicable to try to weaponize this practice, partially with the motivation for financial gain, which we both know you are. Your initial comment is absolutely in bad faith.

 

Separately and additionally, yes, I am exactly willing to take bets once you put up numbers, exactly answering your challenge, which I do not have to do.

2
Agrippa
1y
I think that putting up probabilities is and should be expected. I think that actual financial betting shouldn't be expected but is certainly welcome.  If I was going to dispute the first thing I would do is ask for probabilities. It seems weird to try to argue with you about whether your predictions are wrong if I don't even know what they are. For all I know we have the same predictions and just a different view of other posters' predictions. 

Because you so indicated how interested you are, please proceed by operationalizing this and putting up numbers. I will take bets on either side.

6
Agrippa
1y
In one year we can make a thread that asks EA forum users to vote on whether they believe >90% odds that SBF fraudulently handled funds (that is, in a way that was directly contradictory to public representation of handling funds) in a way that costed FTX.com customers >1bn in losses. If a majority of users (whose accounts existed since yesterday, to prevent shenanigans) vote yes, then the bet resolves YES. Otherwise NO. Which side of 50/50 do you want? 
4
Agrippa
1y
If you don't want to make a single quantifiable prediction on this topic, after making claims about other people's predictions being "too negative", yes I consider that both evasive and inadequate.  If you really believe people are being "too negative" in their speculation, I thought you might be willing to put your money where your mouth is in some way. If you're not, then you're not, but it's got nothing to do with how well defined legality is, the moral meaning of illegality, etcetera. Edit: I don't actually really think that a social expectation of financial betting is a good norm (not that betting is bad, just that declining to financial bet is fine). Please interpret "putting your money where your mouth is" as referring to reputational stake from making a concrete prediction on fraud/malfeasance/etcetera.
2
Charles He
1y
Because you so indicated how interested you are, please proceed by operationalizing this and putting up numbers. I will take bets on either side.

Yes, and no, can you operationalize some of the most interesting and relevant questions to you?

These might touch on:

  • Alameda Research
  • FTX
  • Legal acceptability of FTX's collateralization strategy

 

The fully and honest true answer is sort of no. 

I think legality is poorly posed or not well defined. This is hard to explain and I'm finding the current information environment on the EA forum, and really EA, exceptionally poor and this is disappointing and relevant, if elaborating contributes to the noise.

But basically, whether something is actually found i... (read more)

Lizka
1yModerator Comment31
9
2

The moderation team[1] believes that Charles He has violated a number of Forum norms in the thread below. Because of that and a pattern of previous violations (which led to 3 separate bans in the past — 123), we’re banning Charles He from the Forum for 6 months

Our Forum norms say that we strongly discourage, and may delete, “unnecessary rudeness or offensiveness" and "behavior that interferes with good discourse” — these are the violations we see in the comments below. We appreciate that Charles has unendorsed ... (read more)

-3
Agrippa
1y
Can you just operationalize a few things yourself and attach numbers to them? That sounds easiest.  For example, your odds on whether SBF literally goes to prison within the next 4 years...  (Though I think there are better ways to operationalize) If you can't come up with a way to operationalize a prediction on this topic in any straightforwardly falsifiable way then that's okay I guess, though kind of sad.

It's pretty well done now. This is difficult to follow and this must have been a lot of effort, even exhausting. I think your idea to edit posts was a good one and helpful in this situation.

It's 8 billion now. The ask might be $4 billion.

https://www.wsj.com/livecoverage/stock-market-news-today-11-09-2022/card/ftx-needs-8-billion-bankman-fried-tells-investors-2RSa5oqZyZZ5YrPLuZuL

https://www.reuters.com/technology/ftx-needs-emergency-funding-amid-withdrawal-requests-wsj-2022-11-09/

 

"Crypto exchange Binance reversed course on a rescue offer for FTX Wednesday, leaving the prominent digital firm with an uncertain future as it faces a shortfall of up to $8 billion, according to people familiar with the matter.


Binance chose not to go ahead with... (read more)

The site has been taken private. https://www.alameda-research.com/

 

Sabs — do you think you want to write (negatively) about cryptocurrency in a focused way, as a longer form piece? I think this would be valuable compared to a lot of smaller comments.

1
Sabs
1y
how much are you paying

Note that Hofmann/Semafor is sort of hostile, but as characterized, it's very bad for the prospects of a recovery. 

It's more likely that people don't want to "hold the bags career wise"/work 20 hours a day to fix this for uncertain comp. It's not evidence of conduct (like embezzlement)—I'm pretty sure Hoffman would go for the throat if it was.

Your comment is valid. This reply is getting into sort of low quality/twitter/reddit style speculation and I might stop writing after this:

Some factors that seem different:

  • That was Jan 2021, when I think the economy was still in one of the greatest bull runs in human history, money is tighter now for various reasons
  • At RH, the underlying problem was pretty well defined, and easy to communicate/verify
  • RH probably had more integration/acceptance in conventional finance. Crypto can be openly hostile, SBF was probably relying on less conventional backstops from
... (read more)

Overall, the negative speculation in this thread seems undue and too negative. 

Without trying to make an affirmative statement about what happened at FTX or saying there wasn't any other factors, the comments in this thread ignore the reality of leverage and risk management in brokerage trading (which is what FTX effectively was).

It can be completely true that no customer funds were invested or speculated, but that the fund as a whole can still collapse due to the mechanics/dependencies of leveraged trading.

For example, Robinhood, which no one believe... (read more)

1
Douglas Knight
1y
Robinhood allowed leverage, but I don't think that had anything to do with the crisis.

Would you be open to stating some probabilities on this topic -- for example, your probability that Sam gets convicted of fraud, is conclusively found out to have committed fraud, etcetera? 

I ask because I'd potentially be interested in making some financial bets with you!

Without trying to make an affirmative statement about what happened at FTX or saying there wasn't any other factors, it seems likely that the idea "that customer funds solely belong to the customer and don't mix with other funds" is simplistic and effectively impossible in any leveraged trading system. In reality, what happens is governed by risk management/capital controls, that would almost always blow up in a bank run scenario of the magnitude that happened to FTX. 

For example, Robinhood, which no one believes was speculating on customer funds, had... (read more)

2
David Mathers
1y
Is it not ominous that people were prepared to wriye Robinhood a check, but not FTX?

Well, besides ignoring Nathan Young's huge clout, yes, this seems fair.

Yes you're right. Yesterday morning, the title was less inflammatory and a reasonably factual statement.

The post is less relevant and sliding off the FP, I'll probably delete the entire post at some point (the comments will remain).

This is surprising, like surreal—we are in the simulation sort of vibe. 

I've never seen a website indexed like this in Google scholar. Has anyone else?

6
Erich_Grunewald
1y
Posts from the Alignment Forum have been visible there for some time.

What you said is valid in general. As a reply/rebuttal, I think your focus/association with "lying" is misdirected or stilted.

Basically, what probably happened to FTX is something like the financial engineering described in this link in this top level post: https://dirtybubblemedia.substack.com/p/is-alameda-research-insolvent. This take, which is quite hostile, doesn't involve lying per se.

It's extremely difficult to communicate how normalized something like the above has become, and how much of the "tech" and business success in the 2010s was the result o... (read more)

the quality of the modal post wasn't very good (for its first ~10 revisions), and clearly voted up because of who the authors are; clueful, fair comments are shouted down, while EA leaders making making similar appeals to authority/applause lines  are voted up 

I have no particular opinion on SBF's character, but this paragraph seems ungenerous. Other than being a prolific poster and generally swell guy, Nathan doesn't seem particularly prominent in the EA community.  Certainly not as much as Jonas Vollmer, who's  a key figure at many EA... (read more)

It's almost certain that the residual value after FTX.com's sale will be very low. 

There are also major implications for Alameda trading and the remaining FTX entities. 

Some of the thoughts in this post and thread seem pretty half baked and very uncertain, I think the pace of writing should be lower. 

  • For example, the withdrawals might be at $6B this morning, that would break systems in a purely ops or make movement of money impractical for very innocent, mundane reasons. This experience adds a lot of confusion and noise when reported and echoed.
  • The "Satan's Apple" seems excessively abstract, looking at this from a regular business expansion/portfolio theory seems like more useful and this would benefit from more time.
4[anonymous]1y
What do you mean by this? 

SBF seems like a very good person. Almost regardless of what happened, I don't think a reasonable person's opinion of him should be reduced. 

All of the FTX regranters and the FTX grantees, seem to be very virtuous and effective EAs[1]. FTX funding seemed to be building deeper competency in other areas. This seems like a huge loss for EA.

  1. ^

    I have never received or applied for FTX funding and I never took any actions to lead me being funded by the FTX FF. In the past, FTX regranters approached me.


SBF seems like a very good person. Almost regardless of what happened, I don't think a reasonable person's opinion of him should be reduced. 

I think if he lied publicly about whether FTX's client assets were not invested then I think this should very much reduce a reasonable person's opinion of him. If lying straightforwardly in public does not count against your character, I don't know what else would. 

That said, I don't actually know whether any lying happened here. The real situation seems to be messy, and it's plausible that all of FTX's clie... (read more)

-11
Peter_Layman
1y

This comment is speculative/editorial. 

Coordination / Bailout

Some sort of coordination or saving of FTX seems desirable. Relevant to this, tweets by Liz Hoffman claims FTX reached out this morning privately for aid. The ask may have started at $1B, but in a few hours the need rose to $5B.

https://twitter.com/lizrhoffman/status/1590021299295768578

If this is true, it seems plausible that the smaller amount could have been secured, but the increasingly larger amounts made it difficult. This seems relevant to ideas about coordination or bailout. 

 ... (read more)

8
Charles He
1y
SBF seems like a very good person. Almost regardless of what happened, I don't think a reasonable person's opinion of him should be reduced.  All of the FTX regranters and the FTX grantees, seem to be very virtuous and effective EAs[1]. FTX funding seemed to be building deeper competency in other areas. This seems like a huge loss for EA. 1. ^ I have never received or applied for FTX funding and I never took any actions to lead me being funded by the FTX FF. In the past, FTX regranters approached me.

Honestly both posts are very well written, short and have a lot of content.

What you said seems valid. However, unfortunately, it seems low EV to talk a lot about this subject. Maybe the new EA comms and senior people are paying attention to the issues, and for a number of reasons that seems best in this situation. If that's not adequate, it seems valid to push or ask them about it.

2
Evan_Gaensbauer
1y
I'm thinking of asking people like that about what they're doing but I'm also intending to request feedback from them and others in EA how to communicate related ideas better. I've asked this question to check if there are major factors I might be missing as a prelude to a post with my own views. That'd be high stakes enough that I'd put in the effort to write it better that I didn't put into this question post. I might title it something like "Effective Altruism Should Proactively Help Allied/Aligned Philanthropists Optimize Their Marginal Impact." Other than at the Centre of Effective Altruism, who are the new/senior communications staff it'd be good to contact?

There are writing issues and I'm not sure the net value of the post is positive.

But your view seems ungenerous, ideas in paragraphs like this seem relevant:

This isn't a snide jab at Will MacAskill. He in fact recognized this problem before most and has made the wise choice of not being the CEO of the CEA for a decade now even though he could have kept the job forever if he wanted. 

This is a general problem in EA of many academics having to repeatedly learn they have little to no comparative advantage, if not a comparative disadvantage, in people and o

... (read more)

I understood the heart of the post to be in the first sentence: "what should be of greater importance to effective altruists anyway is how the impacts of all [Musk's] various decisions are, for lack of better terms, high-variance, bordering on volatile." While Evan doesn't provide examples of what decisions he's talking about, I think his point is a valid one: Musk is someone who is exceptionally powerful, increasingly interested in how he can use his power to shape the world, and seemingly operating without the kinds of epistemic guardrails that EA leaders try to operate with. This seems like an important development, if for no other reason that Musk's and EA's paths seem more likely to collide than diverge as time goes on.

Wait!

Unfortunately, this doesn't work, this unjoined me.

 

My statement which discussed "competency" was wrong as written. It was noise without information. Additionally, my original claim did not involve people primarily being in EA-related organizations.

Sorry for the distraction. Thank you for your informative post.

Both the vibes and possibly the reality of tech layoffs/demand for SWE is falling, at least on Hacker news.

https://news.ycombinator.com/item?id=33463908

https://news.ycombinator.com/item?id=33025223

https://news.ycombinator.com/item?id=33083279

https://news.ycombinator.com/item?id=32312303

This seems useful to flag (might be big enough that it seems like an opportunity for the various orgs that are looking for SWE and other talent, even after considering fit/alignment/lemons).

I don't know anything about AI or machine learning. I'm not an EA or longtermist, I just an opportunist/chancer who heard EA has a lot of money so I added AI to my profile.

I don't agree with your post:

  • I think it's unlikely that any practical implementation would provide security, for example, hostile agents could gain access ways by compromising one individual or other pretty prosaic ways, e.g. see Five Eyes. This can be stopped by some very elaborate Signal style service, but that has substantial drawbacks.
  • Even for limited implementations, the nature of r
... (read more)
1
Peter S. Park
1y
Thanks so much for your detailed response, Charles! I really appreciate your honest feedback. It seems likely that our cruxes are the following. I think that (a) we probably cannot predict the precise moment the AGI becomes agentic and/or dangerous, (b) we probably won't have a strong credence that a specific alignment plan will succeed (say, to the degree of solving ELK or interpretability in the short time we have), and (c) AGI takeoff will be slow enough that secrecy can be a key difference-maker in whether we die or not. So, I expect we will have alignment plan numbers 1, 2, 3, and so on. We will try alignment plan 1, but it will probably not succeed (and hopefully we can see signs of it not succeeding early enough that we shut it down and try alignment plan 2). If we can safely empirically iterate, we will find an alignment plan N that works. This is risky and we could very well die (although I think the probability is not unconditionally 100%). This is why I think not building AGI is by far the best strategy. (Corollary: I place a lot of comparative optimism on AI governance and coordination efforts.) The above discussion is conditional on trying to build an aligned AGI. Given these cruxes, I think it is plausible that a Manhattan-Project-esque shift in research norms can help achieve secrecy value in our generated AI safety plans, but also keep ease-of-research high. 

(Taking the post at face value) this is surprising this has so few upvotes . 

  • From this post at least, the OP seems to have a lot of competency (besides getting funding/using resources/networking from EA) and that is rarer these days.
  • For the object level content of the prize/activity itself, this seems useful and could bring in outside resources (not just money) into EA.
3
Håkon Sandbakken
1y
Hi Charles, I must admit I have not followed the international EA movement closely at all the years I've been involved, so I'am a bit confused by your statement. If you have the time, I would like to know what you mean by the statement that it is rare fore EA's to have competency (out side of EA-stuff). Am I to understand that people don't take jobs/work outside of EA-related organizations? My LinkedIn if you are interested in my background: https://www.linkedin.com/in/hsandbakken/ (In Norwegian, but I'm sure you have a translate-function in your browser).

 (To the degree, and on the very substantial assumption that this internet content is valuable and important) this would be useful tag for a hostile AI.

1
Peter S. Park
1y
This would definitely be a useful tag for a hostile AI.  Moreover, the fact that everyone can access tagged posts right now means that there are many ways by which tagged content may still be made accessible by the AGI, even in the event that the AGI corporation cooperates with us. Secrecy is only likely to be kept if the information is known by a small set of people. (e.g., Conjecture's infohazard document discusses this quite well)
2
Charles He
1y
Wait! Unfortunately, this doesn't work, this unjoined me.  
Load more