I think you (your company) and many other commenters here are just trying too hard.
I had just recently lead through several interview rounds for software engineering role and we have not had any issue with LLM use. What we do for the technical interview part is very simple - live whiteboarding design task where we try to identify what the candidate's focus is and might pivot at any time or dig deeper into particular topics. Sometimes, we will even go as detailed as talking about particular algorithms the candidate would use.
In general, I found that this type of interview is the most fun for both sides. The candidates don't feel pressure that they must do the only right thing as there is a lot of room for improvisation; the interviewers don't get bored with repetitive interviews over and over as new candidates come by with different perspectives. Also, there is no room for LLM use because the candidate has to be involved in drawing on the whiteboard and showing their technical presentation skills, which are very important for developers.
Unfortunately, we've noticed that candidates are on another call and their screen is fed by someone else using chatGPT and pasting the responses, as they can hear both the interviewer and the candidate
I saw a pretty impressive cheat tool that could apparently grab the screen from the live share, process text on the screen in response to an obscure keybind and then run it through OCR to solve (or just look up a LC solution).
At that point it seems like trying too hard, but be aware there are theoretical approaches which are extremely hard to detect (the inevitable evolution of sticky notes on the desk, or wall behind the monitor).
I am European, however I have worked with developers from various parts of Asia and South America. English is usually a second language, however most developers are fairly fluent using it as a spoken or written language. Also, most development resources are written in English, so all developers know how to read it. Programming languages and their standard libraries are also written in English. It's the lingua franca worldwide, so we are all happy to use it in the technical context.
We will see by how much. I seriously doubt it improves productivity by a huge margin; if I had to guess, it would be in the ballpark of 10-20%. In the end, it's not the main driver for hiring or lack thereof. Companies that will need to get stuff done and that will be able to make enough money will hire as much as they need to get it done.
It's a baseline which, given the current value of money in mind, in most people's heads defines someone as obscenely/aspirationally wealthy (or even having a multi-generation wealth), yet not someone who could just be endlessly spending this wealth on stupid stuff (having several megayachts, for example) without eventually running out.
Hypothetically, if some government some day sets this amount, then it will pave the way for further adjustments.
But is it really amazing or is it just a natural outcome of countless hours of intelligent humans working together motivated by high pay to further enrich the oligarchs? Would society be worse off without these technically impressive giants? Maybe more competition of lesser products would be better overall.
There have been multiple instances where I would receive invites or messages from obvious bots - users having no history, generic name, sexualised profile photo. I would always report them to Facebook just to receive a reply an hour or a day later that no action has been taken. This means there is no human in the pipeline and probably only the stuff that's not passing their abysmal ML filter goes to the actual people.
I also have a relative who is stuck with their profile being unable to change any contact details, neither email nor password because FB account center doesn't open for them. Again, there is no human support.
BigTech companies must be mandated by law to have the number of live support people working and reachable that is a fixed fraction of their user number. Then, they would have no incentive to inflate their user numbers artificially. As for the moderators, there should also be a strict upper limit on the number of content (content tokens, if you will) they should view during their work day. Then the companies would also be more willing to limit the amount of content on their systems.
Yeah, it's bad business for them but it's a win for the people.
I wonder, should arguments like this be used to justify the issue? With technological advancements things should become better (i.e. safer in this case) without raising inflation-adjusted costs. Wouldn't it be very similar to saying that processors should have become more expensive because they became faster or hard drives because they have larger capacity? Additional costs built into the new codes should decrease over time as builders become more efficient at implementing them or technological advancements allow them to become more efficient.
Not every safety item is necessarily an extra expense due to inefficiency. The simplest example would be electrical wire. 2 lead, paper insulated aluminum wire is less expensive than 3 lead PVC insulated romex, regardless of manufacturing efficiency. There's not a lot of modern AL electrical wire for price comparison, but Home Depot lists a 1000ft spool of #2 AWG USE-2 AL wire from Southwire for $0.80 / foot. The same spool from the same manufacturer in CU is $2.32 / foot. So while not a perfect comparison, if you imagine going from a 1950's 2 wire AL situation to a 2024 3 wire CU situation, you're talking a baseline price increase of over 4x with modern efficiency. And that's before you consider that modern building codes also require many more outlets and circuits than they used to. My home, built in the 70's has 3 circuits in the kitchen. 1 counter top, 1 fridge, 1 range. IIRC the modern building code is 1 circuit for every hard wired appliance, 1 for the fridge, 2 accessory circuits and a circuit for the room walls in general. If you assume a modern house with a fridge, range, dishwasher and over the range microwave/hood combo, that's 7 circuits minimum for the kitchen. So in addition to the 4x increase in wiring costs for the existing circuits, you also have a 2x increase in the number of circuits for just that room let alone outlet spacing requirements in other rooms and the like.
I get these worries; I have them too when hiring potential future colleagues while being a remote employee.
Most people are not fraudsters. Probably you will find them from time to time but it's something that's been disproportionately blown-up by the RTO crowd. There were many people like that in the office - they were forced to show up to work but their productivity has always been non-existent. Signing flexible contracts and allowing the company to fire more easily should prevent vast majority of such hires.
There's another weird point about not being able to detect low-quality work. I fail to see a difference between low-quality work in the office versus remotely. If the employer fails to detect it and pays salary, it's the employer's problem either way.
You are one comment too far with your analogy. Continuing comparing a browser with a child and a corporation with a family is not helping your case.
Analogies work fine for shallow understanding, but like comparing electricity with flow of water inside the pipes, they very quickly break down once you get a bit deeper into the subject.
To make my point clearer, building a browser is not like raising a child, because building a browser is just coding: need more features - hire more developers. They built Chrome not for the good of humanity but for the benefit of their commercial interests.
For whatever reason, Fortune articles appear frequently on HN. Although they almost never receive a high number of votes, I am surprised that this critical-thinking community still has them show up. I understand that Fortune articles are written in a style that appeals to the HN audience (myself included), but they often lack substantive analysis and read more like carefully crafted PR pieces than genuine journalism. The articles tend to follow a formulaic pattern of surface-level business narratives while avoiding any meaningful critical examination, essentially functioning as sophisticated native advertising.
I had just recently lead through several interview rounds for software engineering role and we have not had any issue with LLM use. What we do for the technical interview part is very simple - live whiteboarding design task where we try to identify what the candidate's focus is and might pivot at any time or dig deeper into particular topics. Sometimes, we will even go as detailed as talking about particular algorithms the candidate would use.
In general, I found that this type of interview is the most fun for both sides. The candidates don't feel pressure that they must do the only right thing as there is a lot of room for improvisation; the interviewers don't get bored with repetitive interviews over and over as new candidates come by with different perspectives. Also, there is no room for LLM use because the candidate has to be involved in drawing on the whiteboard and showing their technical presentation skills, which are very important for developers.
reply