As an engineer working on my company’s top of funnel it’s tough. Currently we’ve switched to a short (15-30m) technical problem that we hand grade before candidates get a call. Async technical challenges are obviously gamed but you’d be surprised at how few people both cheat + take longer than 3m to submit the solution
Are you picking a problem candidates can easily cheat on? Could Claude produce the answer in 30 seconds? If so, what's the test actually measuring?
The setups I've seen produce the strongest signal aren't the ones where the candidate can't see the answer that AI produces easily. They're the ones where there is no single answer. Architecture decisions. How you would investigate an outage. How you would tradeoff one constraint against another. The candidate can spend an hour with Claude on it beforehand and it doesn't matter, because the test isn't "did you get the answer," it's "can you talk through the reasoning" and "can I trust your judgement."
Hearing someone explain their thinking on an open problem is a much harder thing to fake. Even if they used an assistant to structure their initial pass, the live unpacking reveals where the thinking is theirs and where it isn't.
> Currently we’ve switched to a short (15-30m) technical problem that we hand grade before candidates get a call.
Funny because I do the same; I don't commit to doing any work/assessments before at least an honest interview round. Businesses have to understand that _they_ are the ones who have to solve for the filtering problem. Most great candidates that you are hoping to snipe are already in a stable job and do not _have_ to jump through hoops to get past a filter.
We’ll skip it based off other signals (referrals notably) of course, but if we offered an interview to every candidate based off their resume we’d never get around to interviewing the good candidates. I don’t like the idea of a technical challenge at the start, but as someone who was discovered way back via skills and not a fancy resume I want to continue offering others the same chance as me.
I have done programming assessments, but I've probably only done 3 in the last 20 years, and those I strictly limited to 1 hour and told the recruiters that I wouldn't spend any longer than that.
I've also flat out refused to proceed with least 5 jobs when recruiters have contacted me, tried to tempt me with a job and then sent through a 3 hour coding test. I've got better things to do with my free time than toy programming puzzles.
As I was reading the start of your argument, I thought you were gonna call the models a depreciating asset! Totally agree about GPUs too, but literally everything they’re spending money on has to be rebuilt to stay competitive. They have to go for the moonshot of training a full new model when better tech comes, they have to upgrade GPUs to keep their data centers efficient.
Technically, the model is a depreciating asset too. Just consider the difference between a model you need a B200 cluster to run vs one you can run on a Raspberry Pi. One's going to have a moat around it that gives it value and the other isn't. It's a hyperbolic argument to be sure but the nature of "enthusiast" hardware is that we're currently running, say, ~27B parameter models on hardware for a few thousand. What's that going to look like in 2 years?
Anthropic/OpenAI really need to train ever-bigger models to keep their moat. But that assumes there isn't a law of diminishing returns and also that a compressed model isn't sufficient for what many people need.
You mihgt say that the training is a barrier. And it is, kind of. Notice how it's Chinese companies coming out with open-source models like DeepSeek and Qwen? That's no accident. As soon as DeepSeek came out I knew what was going on: China is going to make sure no single Western company "owns" AI. It's in their national interest for that not to happen.
I wouldn't be surprised if the rush-to-IPO is motivated, at least in part, by getting ahead of Chinese AI commoditization.
How many failed foundation model training run cycles do you think these companies can tank before the bubble pops and deepseek/etc. catch up to frontier quality?
If Ant, OAI, etc. aren't able to make 20-30% improvements on Opus 4.6 in 2026, does the music stop playing altogether? It seems like they'd lose their ability to charge >10% gross margin on inference in a span of 3-6 months.
As someone who is a maintainer of a few packages, it’s actually really hard to find references for this stuff! The wiki is pretty bare when you start looking for specifics, and like you say there’s tons of crap pkgbuilds when you try to look at what others do.
It helps that the major cell carriers have convinced people to pay multiple times more than the budget carriers, to offer free phone upgrades every few years or when switching carriers. Makes me wonder how much cheaper phone plans could be
The article doesn’t talk about whether immigration enforcement is good, instead focusing on the side effects of the current federal deployment in Minnesota.
Sure, and I'm pointing out what's changed from Obama to Trump is the way that 'sanctuary cities' are complying with federal immigration laws. Under Obama, the sanctuary cities provided lists of students whose parents weren't citizens etc and the media didn't cover the arrests etc. It's selective enforcement of the law. When democrats are in power, they enforce it and media ignores it. When republicans are in power, they defy it and media covers it like it's Nazi Germany. It's absurd and anyone paying attention should reject this framing.
The tragedies unfolding today could have been prevented, with a little human dignity. Speaking of "illegal immigrants" to justify destroying lifes that in some cases had been spent entirely in the US is just that, a cheap blanket to cover inhumane and racist motives. The strawmen you create with "why is no other country allowed to have a border" are as disgusting as your silence to the most pressing current issues. Trump had many options to tighten the grip on immigrants, but i guess this is beyond you aswell.
Please answer me this one. Why is the crime of comming here illegally enough for your kind and nothing else matters? Please argue on a moral basis. I expect the same silence.
I'm not complaining about the past--I'm actually praising Obama for his consistency with uploading the law. You are the one complaining about the present (so that's an interesting rhetorical attack to invert reality like that).
I'm pointing out that democrats complaining about Trump were not complaining about Obama--were you? Or were you one of the ones who took the bait that Trump was building concentration camps (that were actually facilities built under Obama)? You don't think it's funny that the people now calling it fascism were ignoring the same thing under Obama? The sanctuary cities were turning over illegal immigrants who committed crimes to ICE when Obama was in office. They called him Deporter in Chief. It's mass hysteria led by Rachel Maddow and these TDS inflicted people who can't get a grip on reality under Trump. They've truly lost it. Trump isn't Hitler, this isn't fascism, and this is nothing more than sour grapes. In 20 years, people will mock democrats for this period.
> You don't think it's funny that the people now calling it fascism were ignoring the same thing under Obama?
Obamas ICE didnt sweep from door to door.
> this isn't fascism
Except a cleansed bureaucracy, a paralized legislative, and an "totally immune" unleashed executive. Did you get, trump talking about maybe canceling mid terms and invoking the insurection act?
You didnt answer my question about moral, as expected.
DUH because the cities were collaborating and not actively flouting ICE. They didn't need to go door to door. Go back to the same way it was under Obama and this wouldn't be happening.
I asked you spefically, amd reminded you twice now, to morally justify deportations (without due process) simply by crossing the border illegaly -- or better, not having the paperwork. And then i predicted your silence. There is a substancial difference of "crimes" as basis for deportation between trump and previous admins, that alone is a good moral, not legal, justification to not cooperate as a sanctuary mayor.
I also ticked all the boxes of fascism (adimttedly, i left out the ignored judicative branch), to which you didnt respond too.
I dont get how your reading can be that selective. Well, i would strugle too if i had to defend trumps admin.
reply