I think they simply just haven't figured out that the barrier to entry is so low, that no one really cares what their app can do, even if does something genuinely useful.
I've done a ton of low-effort vibe-coded projects that suit my exact use cases. In many cases, I might do a quick Google search, not find an exact match, or find some bloated adware or subscription-ware and not bother going any further.
Claude Code can produce exactly what I want, quickly.
The difference is that I don't really share my projects. People who share them probably haven't realized that code has become cheap, and no one really needs/wants to see them since they can just roll their own.
The kind of code, with the kind of quality, that LLMs can output has become cheap. Learning has not, and neither has genuinely well designed, human designed, code. This might be surprising to the majority of users on HN, but once a really good programmer joins your team, who is both really good, and also uses LLMs to speed up the parts that he or she isn't good at, you really learn how far away vibe coders are from producing something worth using.
I caught myself structuring a comment like an LLM on another site. It's expected that people who chat heavily to LLMs will start to mirror their styles.
I'm surprised how many of my technical team use free ChatGPT in their personal lives. The rest have Claude subscriptions. I'm the only one with ChatGPT and Claude subs and I'll be switching from Claude Pro to Ckaude Max and cancelling ChatGPT, since I only use it when I hit my Claude quota.
I have been out of the loop for a couple of months (vacation). I tried Claude Opus 4.5 at the end of November 2025 with the corporate Github Copilot subscription in Agent mode and it was awful: basically ignoring code and hallucinating.
My team is using it with Claude Code and say it works brilliantly, so I'll be giving it another go.
How much of the value comes from Opus 4.5, how much comes from Claude Code, and how much comes from the combination?
I strongly concur with your second statement. Anything other than agent mode in GH copilot feels useless to me. If I want to engage Opus through GH copilot for planning work, I still use agent mode and just indicate the desired output is whatever.md. I obviously only do this in environments lacking a better tool (Claude Code).
I suspect that's the other thing at play here; many people have only tried Copilot because it's cheap with all the other Microsoft subscriptions many companies have. Copilot frankly is garbage compared to Cursor/Claude, even with the same exact models.
I firmly believe that the heuristics that teachers/lecturers/instructors worldwide use to avoid engaging with reams of mundane text have been successfully by LLMs, and that's why they were so hostile to them initially.
They have to actually read material, and not just use the structure as a proxy for ability.
The relief second officer basically pulled up when the stall protection had been disabled and by the time the other pilot and captain realized what was happening it was too late to save the plane.
There is a design flaw though: the sidesticks in modern Airbus planes are independent, so the other pilot didn’t get any tactile feedback when the second officer was pulling back.
You do get an audible "DUAL INPUT DUAL INPUT" warning and some lights though [1]. It is never allowable to make sidestick inputs unless you are the single designated "pilot flying", but people can sometimes break down under stress of course.
They were well known in fighter pilot community. It's not an obscure fact, they undergo training for ejection seats. There is a specific ejection posture to minimize the risk of snapping your spine.
This is the exact sort of performative garbage that LLMs are great for. I had to do an electrical install, but the installer felt that the code required additional work (I don't think he was trying to rip us off, he sincerely believed it, since it's a volume business model).
I got ChatGPT to come up with some plausible interpretations of the electrical code that allowed the install to continue, including citations. I don't know how accurate it all was, but I sent the argument off to the installer, and he came back and did the work the next day. Even if it gets audited, the chances of the auditor picking apart the arguments are probably slim to none. He has plausible deniability.
This is also why schools and colleges are struggling. No one expected superficially "high quality" work from average and poor students, and now that they have to carefully evaluate everyone's work, they've been caught with their pants down.
Someday superficial AIs will talk to other superficial AIs and they'll deadlock, requiring humans back into the mix. Until then, it's a useful way to do bureaucratic judo.
I get the feeling that you don't know how complicated calculating radiation exposure is. There are plenty of interest in fear mongering against nuclear. Almost all the people talking about how much radiation 300 CPM is have absolutely no idea what they are talking about. Some confuse total measurement units for rates; others are using just the wrong units; still others are talking about levels that are 1000x or 1000000x higher than 300 CPM.
Or to put it another way, 300 CPM (which is a rate) is less than how much radiation you get when on a flight, or how much radiation you get at higher elevations. Even giving a simple explanation of how to calculate Greys (the actual measure you are looking for) takes up the better part of a page. Hell, your bones are radioactive. Yet there are plenty of people posting that somehow the risk to this guy is radioactivity. In reality, his biggest problem is probably going to be finding a new job.
CPM doesn't mean anything, it's useless to compare it to anything without more information.
For his job, depends on the dose he took. In my country he would have been on benefits until the dose was calculated, then if possible, reintegrated in the team, or directed towards a new job if not (paid formation and everything). I've studied with a diver who couldn't work with radioactive trash anymore, he wasn't meant to be a SWE in the end, he now dive for unexploded WW2 stuff in the north sea/Baltic I think
In a way you are right, but in another way you are wrong. Background radiation is mostly gamma which is generally the most dangerous kind. Radiation at a nuclear plant is mostly alpha and beta which are less dangerous at the same CPM rates. So technically you are right, you have to calculate the absorption, but in practice you are wrong because at that rate, there is no way he absorbed enough radiation to be dangerous.
Yeah I stopped really looking at replies when I realized it was just a bunch of people telling me that falling in a nuclear reactor pool isn’t that bad and to go read the XKCD again if I don’t get it.
reply