onion2k

10 ms·

211.
by onion2k 3mo ago
There's a 'joke' that goes around occasionally that has some truth to it: "Excel is the world's most popular programming language." Occasionally it's 'Excel macros' or 'VBA' instead of
212.
by onion2k 3mo ago
It's also possible that giving access to Fable when lots of people have been taking time away from working on things (4th July, World Cup, etc) meant they didn't spend much time with it.
213.
by onion2k 3mo ago
If thorough testing and 100 experts can’t find a problem, the thing is probably perfect. If you can get 100 experts to agree on something then you've cracked a much harder problem than software quality.
214.
by onion2k 3mo ago
Unless you are extremely crafty, you don’t get to retire from this sort of criminality. He got away with it for 30 years. That shows at least some level of craftiness.
215.
by onion2k 3mo ago
I don't really understand the mentality of people who do this sort of criminal activity. If he'd stopped after, say, $5m and just retired he'd probably have managed to get away with it. Continuing to such a ridiculous degree
216.
by onion2k 3mo ago
If a fancy new feature can’t degrade gracefully, then 98% isn’t “widely supported”. Close, but the other way round. Don't avoid a feature because it lacks good enough support. Write code to progressively enhance the experience if the
217.
by onion2k 4mo ago
Unlike all your examples, switching out an LLM is both cheap an easy. Rolling out AI access in a large business is still hard, especially if you're trying to do it safely e.g stopping people throwing all your company data including u
218.
by onion2k 4mo ago
New colleagues can be taught that, LLMs not so. LLMs can be 'taught' though. You can give them additional context or instructions. The difference is that they can't really teach themselves. This is roughly what I'm say
219.
by onion2k 4mo ago
OpenAI and Anthropic aren't charities, so whatever cost they inccur for training will be passed down to the companies using the models You should, but with two important caveats. First, you don't know what their amortization sch
220.
by onion2k 4mo ago
The bit I don't have evidence for is whether or not teaching ICs to be managers would improve how they use agentic AI. I have plenty of evidence for the efficacy and effectiveness of AI itself (although not qualitative or obviously cau
221.
by onion2k 4mo ago
Managers don't really read or review the codes. Good ones do. Reading the code someone is contributing is a powerful signal about how well they're doing. ICs who manage swarms of agents should operate the same way. They set them
222.
by onion2k 4mo ago
Do I write all my code with an agent now? Yes. Can you just give an agent a desired outcome and let it work, unsupervised? Absolutely not. I strongly suspect that developers moving from writing code to managing agents to write code for th
223.
by onion2k 4mo ago
Totally happy with them. I don't use any of them, and that's working really well for me.
224.
by onion2k 4mo ago
If I keep running the same prompt, it will keep finding issues. I've had the same experience, but whenever I've reviewed what it finds it's basically right. It's pedantic, and a lot of the problems aren't things I
225.
by onion2k 4mo ago
I don't think the "it's piracy to use the code generated by an LLM because it closely resembles the code the LLM was trained on" argument has been fought in court yet.
226.
by onion2k 4mo ago
Google says that James Webb telescope cost a total of $10bn. That's not a lot of money in the grand scheme of things. Private citizens could afford to put similar things into space if they chose to. We don't need them to be cheape
227.
by onion2k 4mo ago
The page says Playwright tests " breaks on selector/layout changes " but that this is more robust because it " reads the page like a human, via ARIA semantics ". Any well-written e2e test should be using ARIA semant
228.
by onion2k 4mo ago
If you were to say 'doing the same work a bit faster' isn't life-changing then that's evidence that the hypothesis in the article is actually correct.
229.
by onion2k 4mo ago
small downsides like security holes? Those aren't small I'm British. I've been taught to turn understatement into an art.
230.
by onion2k 4mo ago
I would, but that's mostly because I don't trust PRs to catch real problems. Someone reviewing the changes in a codebase is never going to spot an architectural or code design issue, and those are the real ones you need to care
231.
by onion2k 4mo ago
Generation got cheap. Verification got expensive. That proves AI is capable of doing one part of the software engineering process. The 16 devs in the study trusted AI to write the code. Once we trust AI to do the verification as well we&#
232.
by onion2k 4mo ago
If you get enough signals like that you can often narrow down a very large cohort of people to an individual. First it's 'over 18?', then it's 'over 25?', and then 'biological sex?', 'employed?
233.
by onion2k 4mo ago
I don't. I write a lightweight API proxy, usually as a CLI app, that takes requests from the agent, adds secrets and makes the request, and then passes the response transparently to the agent. The agent can use APIs without ever seeing
234.
by onion2k 4mo ago
Ho-Ly - LHC maybe.
235.
by onion2k 4mo ago
If I did have fully pure Pu-241, would that be a better fuel than U-235? Also not a physicist, but I assume from the fact that the OP is asking the LLM this question to trip it up, the point is that U-235 is better even if you have an ab
236.
by onion2k 4mo ago
There will always be industries that require auditable code, so some things will need to be human-readable until we trust AI to audit code that's written by AI. I imagine that's a long way off, simply because no one will trust p
237.
by onion2k 4mo ago
None of the examples reflect 'real work', at least not what I'd consider real work. Being able to nail a zero-shot greenfield project is relatively easy even for a small model. There's not much context to build up and
238.
by onion2k 4mo ago
it's also unreasonable for me to expect that Apple should be hosting my 99¢ purchase for the rest of time That's fair so long as it was made clear at the time of purchase that you aren't buying the song, and that you'r
239.
by onion2k 4mo ago
My partner has been a subscriber for years. It's great. Stewart Lee's music reviews have been a highlight, but there's great long form content in most editions.
240.
by onion2k 4mo ago
Sorry, this is going to be a rant. :) Admittedly it's a long time since I was in school, but when I was there the notion of teaching the benefits of physical activity was limited to being told to run about for some 'sport', a

More ›