There will have been tests, but there will have been missing end-to-end tests. Test 1 will verify that the new system/product emits billing entries in some expected way ("We did 100 bytes of operations and we see we called the billing system for 100 bytes of stuff, yay, test pass"). Test 2 will be in the billing system ("We provide an incoming bill for SKU#12345 for 100 gigabyte-units and we see it costs $17, yay, test passes"). But they won't test the two things together because it will be harder to do and the teams will have different management chains. Seen it happen several times at several companies. Somebody will have said at some point "we should actually have the tests charge money" and somebody else will have said "well we can't have the tests actually charge money, that's a legal/accounting problem, it might even be a crime" and then nobody would have asked what the next best thing was.
Another tactical move is to just stop. You're allowed to exit the AI business. Nobody's forcing you to keep throwing money into the furnace. Just be a rocket company. All of the xAI founders left. Your product's brand name is mud. Just stop doing that and build spaceships.
> The Paxos algorithm for implementing a fault-tolerant distributed system
has been regarded as difficult to understand, perhaps because the original
presentation was Greek to many readers.
Ha! That's very clever, author. You clearly have a similar sense of humor to...oh, it's Leslie Lamport again.
I interpret "keeps rising" negatively. Changes keep getting made, certainly. The AIs will perhaps never fail to fulfill your feature request. But there's no overall plan. It's just undirected, cancerous growth. It's Homer Simpson telling a team of automotive engineers to add feature after feature.
Heck, even before they do anything, there is a sizable industry (with just a few players) focused entirely on the incredibly byzantine (but originally well meaning) process of bidding for government contracts (and also the politicking of acquiring no-bid contracts).
Amazing character. Started as a regular robot-loving engineering kid, was in the right place at the right time and earned something like $140 million from Google, mostly from truly ludicrous performance bonuses, went to Uber for another giant payout, was worth nine figures. And sure, he was convicted for crimes, but he got one of those definitely-legitimate Trump pardons.
And then he managed to turn that into a negative $50 million net worth.
And also he briefly started a religion based around having an AI inventing a Christian god or something because his story wasn't crazy enough.
Also, I love that it was open sourced. Although it sounds like from the GitHub page summary that there was some shenanigans involving a "certain very high profile game studio" that I'd love to hear more tea about.
> They select for comfort with the prevailing mess, because they have no other frame of reference.
This doesn't go far enough. Employees on hiring committees select for conformance with their peers on the hiring committee because conformance with the other hiring committee members is the only success signal they will ever receive. An interviewer at a sufficiently large company will never receive any feedback (let alone timely feedback) on whether they are choosing good or bad coworker candidates. They will only ever get feedback on how other interviewers vote. In such an environment, the best you can learn to do is to conform.
In a small startup, where you immediately begin to work with the person you chose, you will get a lot more feedback on the people you chose to hire. Even then, though, you will never get any feedback on your false negatives. Are you rejecting lots of good candidates? You'll never know.
> GAO is a Congressional agency, it does not fall under the Executive
Ah ah ah, you're describing how things were before Trump v. Slaughter, when the Supreme Court justices ruled that Republican Presidents are allowed to fire the heads of non-executive agencies so long as they are not the Federal Reserve.
Let's try the suggested advice and ask why five times.
> Why do you even need an AI assistant here?
To take notes on what we talked about.
> Why do you want to do that?
Because I want to retain the contents of this conversation, and I don't want to be distracted by note-taking. I want to be in the moment and also have a record of what we talked about for later.
> Why do you want a record?
Because I expect that what you're going to say is valuable enough to want to reference later. Perhaps you will give me the name of a cool podcast, or you'll give me very good, detailed advice. Perhaps you'll mention when your upcoming birthday is or a favorite brand of a product, and it'll be useful for gifting.
It has occurred to me over and over for the last several years that many of the senior engineers at my company would be substantially more productive with some sort of assistance from someone who specializes in having executive function skills. One such person given four or five engineers to manage could do wonders. And they'd also be the best source of feedback for hard-to-measure performance evaluation information like "is this senior engineer actually working on anything most days?" But executive assistants are a privilege reserved for only the people who are at level X and up.
Yeah, Amazon has disadvantages (counterfeits, fake reviews, more expensive) but its advantages (insane shipping speed/cost, strong return policy) are nearly impossible for competitors to compete with. The manufacturer is not going to offer same day delivery and no-questions-asked returns.
Honestly, given the dangerously unmitigated power of Claude Mythos, we should really look into arresting the people who have failed to ask Claude to cure cancer already.
> randomly filtering "too many" resumes is pretty much allowed (I think)
It's totally fine to filter out resumes in a completely random, content-independent way. Grabbing the fourth resume down in the pile and offering them the job is a perfectly fair albeit stupid way to make a hiring decision. However, AIs are very, very good at capturing biases, and it would not at all surprise me if an AI told to filter resumes is going to end up filtering with some biases for things that you definitely do not want to filter on, like the name of the candidate. And it might be that everybody resume that claims it fixed a typo in a major open source project gets a pass, but resumes that only list their own projects get rejected 60% of the time, so you're losing more good candidates than bad.