I really like this style of testing -- code that can be tested this way is also the most fun kind of code to work with and the most likely to behave predictably.
Per-minute pricing for self-hosted runners seems like a very fast way for them to force everyone who actually is using self-hosted runners to migrate away.
I suspect we'll be doing that sometime in January or February.
I'm more and more convinced of the importance of this.
There is a very interesting thing happening right now where the "llm over promisers" are incentivized to over promise for all the normal reasons -- but ALSO to create the perception that the "next/soon" breakthrough is only going to be applicable when run on huge cloud infra such that running locally is never going to be all that useful ... I tend to think that will prove wildly wrong and that we will very soon arrive at a world where state of art LLM workloads should be expected to be massively more efficiently runnable than they currently are -- to the point of not even being the bottleneck of the workflows that use these components. Additionally these workloads will be viable to run locally on common current_year consumer level hardware ...
"llm is about to be general intelligence and sufficient llm can never run locally" is a highly highly temporary state that should soon be falsifiable imo. I don't think the llm part of the "ai computation" will be the perf bottleneck for long.
If you succeed in creating a generic async primitive, it doesn't really matter what the original task was (as long as it's something that requires async), no? That's an implication of it being generic?
It takes more than just misaligned incentives to get a banking crisis -- you have to have structural corruption preventing the transfer of
the loss gradient back to the "misaligned" decision makers. It's somewhat disingenuous (or overly innocent) to reimagine the pathways which power structural corruption as "innocent ignorance in the face of bad incentives".
The real world has "actually bad" actors -- not just misaligned incentives.
> It takes more than just misaligned incentives to get a banking crisis -- you have to have structural corruption preventing the transfer of the loss gradient back to the "misaligned" decision makers.
Nah, you can do it just on the basis of information asymmetries.
Banks can sell mortgages. People think buying mortgages is safe, because banks don't loan money to people they don't think can pay it back, and even if they did, the mortgage is backed by the house so in the worst case you can foreclose and get back your principal. So lots of people buy mortgages.
Then banks figure out that it's easy to sell mortgages, and that if they sell them it doesn't matter that much if the people they loan the money to can pay it back. Plus, the less creditworthy people pay higher interest rates, and you can still foreclose if they default. So banks make a lot of loans to people who can't afford them, and then sell the mortgages, and people still buy them.
Except that if this happens at scale, the people taking out mortgages they can't afford bid up the price of houses. And then when they start to default and you want to foreclose, you'd have to sell the house to get back the money, which at scale means that the prices would go back down to where they were before they got bid up, which means you wouldn't even recover your principal.
If everybody realizes that this is what's going to happen then people wouldn't buy bad mortgages from banks and then banks wouldn't issue them. But if enough people don't notice until after the bubble is inflated...
You can sit them down and explain precisely why buying something, like a new car, is a bad financial decision and that they cannot afford it anyway, and then watch them go buy it anyway. To the point where I have seen people laugh about how dumb of an idea it is, while in the act of doing it.
The "I wish someone explained to me..." that comes later when it all falls apart is largely just licking the wounds of their damaged ego.
> You can sit them down and explain precisely why buying something, like a new car, is a bad financial decision and that they cannot afford it anyway, and then watch them go buy it anyway. To the point where I have seen people laugh about how dumb of an idea it is, while in the act of doing it.
And this is actually fine because it comes with its own integrated stupidity penalty. We only need the government to impose a penalty if the person who needs the disincentive when making a decision is different than the person being affected by it.
This is a big "hell yes" for me! Some seem to think that mortgaging themselves up to their eyebrows with huge houses and the latest vehicles is a good idea. As an example: I needed a pickup truck back in 2021. I settled on a ram and purchased the base model. The only options were a towing package and the medium level smart audio/display system for a cost of $27K. I could have easily spent $50K and got a whole lot of other options, but determined the extra cost was too much and the options weren't needed. (The only reason I purchase a new one is people tend to drive like maniacs in trucks where I live, so I didn't trust a used one.)
I digress, the numbers alone are the reason for the base model, because I could use the extra money somewhere else. And yes, new vehicles do depreciate too much. However, if you keep the vehicle for it's entire lifespan, the hit isn't so bad.
You skip over a very important step here, where people keep buying the MBSes because the ratings agencies are knowingly rating the securities incorrectly. If that didn't happen, the market would be too small to blow up in the way that it did, all of the safe money can't invest if the MBSes aren't AAA.
It's not that no-one noticed in time, it's that the people responsible for noticing were paid to pretend they hadn't. That is the corrupt part.
What they were doing was, they'd take a bucket of high risk mortgages and apply a contract to them to retroactively sort them. So, if you bought the 30th percentile of the bucket and then anything more than 70% of the people in the bucket paid their mortgages you would get paid, and if fewer than that did then you wouldn't.
Then they were rating the highest percentiles in the bucket as AAA because even for borrowers with bad credit, the probability that such a high percentage of them would default was considered very low. Even for people with bad credit, default rates are usually only something like 10%.
But that doesn't work out if you haven't noticed that banks have stopped caring about the default rate when issuing mortgages.
I disagree with the characterization of structural corruption. Every rationale actor will seek to capture all the benefits and pass on the risks. The real corruption is when decision makers know that they can’t be held responsible through corporate or political structures. See also [moral hazard](https://en.m.wikipedia.org/wiki/Moral_hazard)
In our ci setting up the docker buildx driver to use the artifact registry pull through cache involves (apparently) an auth transaction to dockerhub which fails out
The only figures of note that were assassinated that i can think of were more lefty -- or at least non right -- jfk, mlk, harvey milk, bobby kennedy, malcolm x -- were there actually
any prominent american right wing figures assassinated in this "period of escalated political assassinations ...?"
I love determinism and plain old data.
reply