There has been at least one posted here in Hacker News, Mojo. Google shows some other similar attempts.
The real issue with doing this is that there is no body of code available to train your models on. As a result the first few look like opinionated Python.
I tried this and found that there are a set of potentially good candidates who simply cannot bring themselves to criticize anything in front of an interviewer due to the power dynamics of the interviewer/interviewee relationship.
The iterative approach described here to finding a 'good' borgizaf is simulated annealing. Make an initial guess to the solution, then start with changes that are large early on and which increasingly refine as you get closer to your goal.
Metropolis-Hastings algorithm works better for situations for which you know even less about what a good borgizaf looks like. Perform many experiments (code changes) and toss out all of the ones that don't work. Science works this way. An automated LLM coder would probably also go this route as the cost of trying many things would be more heavily automated.
Large systems tend to look more like the genetic algorithm. The space is diced up into individual components and then each 'gene' is optimized in parallel. For example if you were trying to build a Linux distribution you'd have several hundred packages and then for each release the packages could improve or be entirely replaced with better versions (or swap back and forth as they competed).
Of course there are other search strategies that can be employed. Search is still an important area of research.
Sure, everything is an optimization problem, the hard part is defining your cost function, especially if the borgizaf is trying to solve an ill-formed business problem.
This is a legal thing. If you do a layoff it's for business reasons and you can avoid all of the PIP and such. But if you do it that way you can't select based upon performance.
Isn’t this why some people are so into performative work? In a layoff the people they suspect might be underperforming go onto the list. They keep the people who look good on paper, the ones who play the game.
Not the “untapped” people the author is talking about.
That's exactly it. The untapped people are actually getting bunched into the "underperforming" category because in the eyes of the beancounter they are not meeting some benign performance metric that the company wants to see.
Say I'm a phone support company. I have a script I want my employees to follow and the average support time per phone call should be anywhere between 15-30 minutes. Sally Sue is on the phone for the full 8 hours and handles 16 calls a day. Billy Brass is on the phone for 4 hours of the day but handles double the amount of calls a day.
To the bean counters Billy is underperforming because he only spends 4 hours time on the phone and the company only makes money for the amount of time they can keep people on the phone. In this example it doesn't matter that Billy is an all-star because he completed more calls, he's underperforming because he's not following the script that should keep people on the phone for as long as possible.
The point is that Billy will feel resentful because even though he's able to help more people in less time he's getting penalized so Billy has less incentive to go above and beyond and in fact needs to degrade his workflow to fit someone else's metrics. So Billy becomes "untapped" because the company has restricted his autonomy. He "CAN" do more but that's not what the company wants from him so he will choose not to do it even if it's to the benefit of the company.
It would probably work if it were just a -$1 ad credit on each download? Except it looks like they are making their money here on sponsorship and upsells rather than youtube ad revenue.
Not sure about trains, but large ocean-faring nuclear-powered tugboats have been considered to move container ship sized barges most of the way and then let battery or fossil powered tugs bring em the final mile into port.
The cost of labor is too low, and the prevalence of pirates too high, to justify this idea. You're also talking about serious intermodal requirements to handle these things when they get to the interface between open ocean and restricted waters, because safe navigation is a near impossibility with the tug configuration you're proposing, especially once navigational constraints (e.g. land or shoal water) come into the picture. A large merchant vessel takes several miles to stop on its own, before you consider any towing configuration. If you've never operated around them, you probably don't have the instinct that you can be in serious danger of a collision while still several miles away from another vessel, but what you're proposing is scary.
Even without the navigational problems, you're necessitating armed guards by having these ships operate internationally, because these ships would be huge pirate targets on a much more severe level than the scrap metal and parts theft that is present throughout the world. That gets... tricky, legally (see https://www.swedishclub.com/upload/Loss_Prev_Docs/Piracy/PIR...)
Also, cargo ships are already enormous. Trains pale in comparison. We don't need these headaches. Oceangoing shipping is miraculously cheap already.
Edit: I'll also point out that, regarding the sibling comment's link, oceangoing shipping was a much smaller club in 1963, so the navigational challenges would not have seemed so daunting. Containerized shipping didn't even exist until the mid-50s, and AIS and radar let us safely pack way more vessels into small spaces than they could have suspected.
Pirates and the habit of shipowners to hide responsibilities in a maze of sub contractors, charterers, and flags of convenience.
The the maritime industry's track record of preventing oil spills is pretty bad. Do we really want those people to routinely operate nuclear reactors?
This industry has an issue with faked crew certificates (https://www.hellenicshippingnews.com/fake-certificates-conti...) "An EU-sponsored pilot project, branded ‘GetQuality’, revealed that “nearly every tenth seafarer worldwide” has experienced fraudulent certificates"
Insurance is already a tricky issue for land-based nuclear power plants. Who would insure a nuclear-powered ship? This was a major issue for the German nuclear cargo ship Otto Hahn.
The feasibility of convincing enough countries/port authorities to permit you park a nuclear reactor in their port is fairly low. If something goes wrong, they'll be the ones that have to clean up the mess after your shipping corp declares bankruptcy or otherwise skips out on the cleanup bill.
Some countries do mind even then (New Zealand particularly.)
But it's probably easier to believe the US would stick around to help clean up the mess, since the US Navy will presumably continue to exist and desire to operate in foreign ports in the future. A private shipping company on the other hand could effectively poof out of existence once they become liable for a nuclear accident cleanup.
Corn will not pollinate correctly if it is not in a rather dense clump. If you grow it in your garden you should plant 3 or more rows(or a square/circle). Otherwise you will get mostly empty cobs.
3 rows is not nearly enough. I grow corn, and even with 9 rows, 20' each (roughly a 20'x20' square, with nearly 200 plants), I still hand polinate (takes about 5-10 minutes every 2-3 days once you start seeing silk). I doubt that most home growers will get anywhere close to enough density for good wind pollination, but hand pollination is not that hard.
The real issue with doing this is that there is no body of code available to train your models on. As a result the first few look like opinionated Python.