Well, that's kinda the problem, isn't it? Even after being erroneously charged and ghosted by their non-existent support for a month, you'll still happily keep paying for their services.
If most people think like you, why indeed bother providing support at all?
Good point. I did actually cancel my Claude subscription a week or two ago, but I renewed it (regretfully) just the other day. The only other SOTA model that seems to be on-par with Opus 4.6 for engineering work is (maybe?) Codex 5.3, though I would rather not support Sam Altman indirectly.
Unless I'm misunderstanding something, this isn't that big of a number in the larger scale of US banking; According to the numbers in the article that's only about 2.5% of all bank lending (300B/1.2T, with the 1.2T being ~10%)
> this isn't that big of a number in the larger scale of US banking
It's not. It's just that we're seeing potentially 10% losses on the portfolio level [1], which could imply up to–up to!–5% losses to the banks' loans to those lenders.
Again, tens of billions of dollars of losses are totally absorbable. But Morgan Stanley's stock price took a hit when it gated one of these funds [2]. And some banks (Deutsche Bank, somehow, fucking again, Deutsche Bank) have small ($12n) but concentrated portfolios where a single wipeout could materially impair their ~$80bn of risk-weighted assets.
> Again, tens of billions of dollars of losses are totally absorbable.
They are, in isolation. The _problem_ is that PE doesn't generally trade assets in public, which means that valuation only really come when you're either wanting to buy, wanting to sell, wanting to re-loan or in deep shit.
But! banking can absorb a few billion right? yes, so long as people are not asking questions about other assets.
Because PE assets are not publicly traded (hence private in private equity) the value of assets are calculated at much lower rates than on a public market. This means that the assets that PE holds could be wildly over or under valued. The way we assess the value of PE holdings is thier looking at the Net Asset Value calculations (which might be done twice a year) or infer the value based on public information.
Now we are told that markets are rational and great at working the value of things. This dear reader is bollocks. Because PE is a black box, if a class of asset that they hold (ie SaaS buisnesses, or high street stores, or coffee trading etc) looks like its not doing well, people will start to write down the value of people holding loans given to PE, or shares in PE.
This creates contagion, because one PE company is in distress, the market goes "oh shit, the whole thing is on fire" and you get bank runs (because where is the money coming from to loan to PE? thats right banks, eventually)
Washington Mutual had $307 billion in assets, and one credit downgrade and a bank run of $16 billion in September 2008 was enough to get them shut down.
These private credit numbers are estimates provided by Moody's, who were famously clueless about the scale of mortgage bond risk even as they stamped them all with a AAA rating.
Someone else owns all the other credit. This is the 1st domino.
The liquidity challenges of a $1.2T shock to the economy is meaningful, because it has knock on effects on equity as well.
When private credit (which is propping up private valuation) falls, private equity also falls and then everyone realizes that everyone else has been swimming naked.
Yeah but don’t major problems usually take the form of debt chains being forcibly unwound? Like default happens somewhere, expected money doesn’t land, that next entity who was receiving the debt money that defaulted defaults on something because now they can’t pay, expected debt payment money from them doesn’t land, that next entity etc.
So I think it’s not about how much of the debt is this, it would be about how intertwined it is with other things.
I’m not claiming it’s major btw, I’m just clarifying that in my understanding it could be a small percent but still end up causing default cascades, or it could be a large percent and not, depending on the debt graph.
Update: original comment should be. 300B/1.2T*(10% of bank funds) = 2.5%. If I'm reading comment correct. Also I believe the whole private credit ecosystem is about 1T.
In a catastrophic scenario: if the whole asset class went to 0 (on the banks asset sheet they would lose 2.5% - absorbable pain assuming its not leveraged through creative financial mechanisms).
I would wager that risk is more concentrated on certain institutions instead of across the board so acute pain likely.
I've been told by the head of compliance of the largest European banking group that 2.5% is exactly the threshold at which they begin to be very worried/ at systemic risk
Apparently they operate on very low level of tolerable risk (way lower than I thought)
>2.5% is likely still survivable, but i think risk departments + regulators are all a lot less risk tolerant after seeing how quickly things went south in 2008 and worries about an out of control spiral
It matters to me because I'm reading it now and feel more informed about this problem. Throwing the towel in and saying it's all pointless isn't helpful.
It's not throwing in the towel, it's about doing things that we the people can actually do.
One thing, we the people can do, is pressure our politicians to break up Google along with the rest of big tech.
There are many primary challengers this cycle that are running anti-monopoly platforms. Help their cause, signing pointless petitions is just West Wing style fantasy that is extremely childish.
Because the company either has to address it, or stop pretending it's "listening to concerns" or whatever. Even if it doesn't change the outcome, it makes it clearer that the company is engaging in bad faith.
It's something apps that will soon break can point their users to so they know to blame Google and a bunch of incompetent governments.
Google will not change their minds, they're too busy buying goodwill from governments by playing along. There aren't any real alternatives to Android that are less closed off and they know it.
Unless the AI industry can figure out how to be profitable soon (to which basically nobody has a clear path to profitability besides maybe ad revenue) it's hard to see this not blowing up in a year or two. The bills for all of this are going to come due eventually and the AI CEOs can only convince investors to keep letting them burn billions for so long.
It's incredible how much money is being lit on fire without anyone having an answer to the question "how will you generate revenue". Meanwhile alternative energy discussions here are met with harsh criticisms of their bottom lines while they have some of the most firm demand of any industry. Society should be prudent with their investments, but only when it's used for R&D into essential stuff and not a bunch of toys for rich psychopaths?
"If we end up misspending a couple of hundred billion dollars, I think that is going to be very unfortunate, obviously. But I actually think the risk is higher on the other side . If you build too slowly and superintelligence is possible in 3 years, but you built it out assuming it is possible in 5 years, then you are out of position on the most important technology."
His assumption is that superinteligence is close, its just a question of whether it is 3 or 5 years!
No, the first sentence makes it clear that he is aware AGI is not a given. His position is that there is a possibility we can reach super intelligence, and given that possibility they want to be on the bleeding edge and are investing accordingly, given that even total failure won’t cripple their business and all their competitors are doing the same.
"If you build too slowly and superintelligence is possible in 3 years, but you built it out assuming it is possible in 5 years, then you are out of position on the most important technology"
This made me think Zuck sees it as a question of when rather than if. I.e its more a question of 3 vs 5 years rather than possible vs non possible.
And yet, imagine it's 100 million jobs, at 100K per job. That's 10 trillion dollars a year, well worth the investment! Except that it won't be 100K per job that AI companies will capture, it'll be 2K. So that's not 10 trillion dollars a year, it's 200 billion. Which is about what a big tech company makes in advertising already.
Speedhunters was the best way to experience broad car culture as a young kid in high school, in its prime it used to be two articles a day of well-written and photographed car features from all over the world in every style, all presented with no ads. A shame to see it go as there's nothing else really like it out there, now it's all scattered across instagram and various youtube channels. I'm hoping EA sells it off to someone who cares but knowing their track record they're likely to completely shut it down as soon as someone remembers it exists
reply