2
points
Zababa
Joined 5,705 karma
- ZababaBeautiful acknowledgment list, and congratulations on the beta release!
- Quoting the article:
>I’m having to choose my words carefully, because I need to stress one thing: these are not the most common reasons for men and women to be admitted to hospital. They are the most typically male and typically female.
If you go to https://leobenedictus.substack.com/p/that-hospital-admission... and sort by number of admissions, you get stuff like:
- Personal history of certain other diseases
- Personal history of medical treatment
- Personal history of allergy to drugs, medicaments and biological substances
- Personal history of other diseases and conditions
- >For kids that was a way to decide who is going to fetch the water for the table (smaller number or higher number of the table).
I only knew the version where your age is the number on the glass. For fetching water, it was the slowest person to say "pot d'eau" (water jug) and sometimes put a hand to your head (it depended on the group).
- Go has iterators, had them for a while now. To delete an element from a slice you can use `slices.Delete`.
>3) To make matters worse, at work we have a linter that forbids merging a branch if you a) don't do if err != nil for every case b) have >20 for & if/else clauses. This makes you split functions in many pieces, turning your code into enterprise Java.
That is not a problem with Go.
- You're really good at writing. Best of luck
- There was a good series interviewing people that worked in both software engineering and traditional engineering: https://www.hillelwayne.com/post/are-we-really-engineers/. The conclusion was that yes, a lot of what we do as software engineers is engineering.
- Yeah it is genuinely terrible and getting worse. It's kinda fine if you have your own little bubble and can adapt to frequent algo changes and only use the "following" tab, but then you miss on sometimes pretty good recommendations. I used to tell people to get on there to have the latest AI news by a few good accounts, I don't anymore, I don't want to expose them to all of that hate.
- Btw the definition Karpathy gave was:
> a system you could go to that can do any economically valuable task at human performance or better.
https://open.substack.com/pub/dwarkesh/p/andrej-karpathy?sel...
- > Amazon’s data centres were projected to use 7.7 billion gallons of water a year by 2030, according to the leaked strategy memo, which was circulated within the company in 2022.
From https://www.usgs.gov/water-science-school/science/total-wate...:
> Water use in the United States in 2015 was estimated to be about 322 billion gallons per day (Bgal/d), which was 9 percent less than in 2010.
It doesn't seem to be very much water at all.
- > The engineer, whose name was Craig Hannah, was also a keen naturalist and photographer. He saw the same thing happening repeatedly and wondered if it would be of interest to insect researchers. This led him to the University of Exeter’s Centre for Ecology and Conservation, to which we are both affiliated.
> Craig diligently collected small specimen-tubes of flies at the rig, which is in the UK Britannia oil field, and they started arriving regularly on our desks. We’ve spent the past few years studying them, and the results have now been published for the first time.
I really like it when this kind of thing happens. Someone being curious, contacting experts, experts being receptive and working with that person.
Edit: this may be this person's flickr: https://www.flickr.com/photos/21913923@N03/53747119426/in/al.... A few of the oil rig, a hoverfly, and generally lots of beautiful pictures.
- Doesn't fully correct for different biaises like healthy user bias so it proves association more than causality.
- > LLMs won’t get much better, but that is okay
> The last releases were unimpressive. Does anyone know a real application where ChatGPT 5 can do something that o3 could not?
Yeah, GPT-5 is way better at coding in codex, especially for longer tasks. Opus 4.1 is pretty good too. Gemini 3 is dropping soon. GPT-5 was more about reducing costs/having a router so when your doctor asks a question to chatgpt he's routed to the thinking version from what I understand.
> The good news is that what we have is enough for most people.
This is true.
Also stop using LMArena as an indicator of anything, it hasn't meant much for more than 6 months.
- Yeah that's a good point and the sibling comment seems to be pointing in the same direction. You could take a look at Steve Yegge's beads (https://steve-yegge.medium.com/introducing-beads-a-coding-ag..., https://github.com/steveyegge/beads) but the writeup is not super detailed.
I think your last point is pretty important, that all that we see is done by experienced people, and that today we don't have a good way to teaching "how to effectively use AI agents" other than saying to people "use them a lot, apply software engineering best practices like testing". That is a big issue, compounded because that stuff is new, there are lots of different tools, and they evolve all the time. I don't have a better answer here than "many programmers that I respect have tried using those tools and are sticking with it rather than going back" (with exceptions, like Karpathy's nanochat), and "the best way to learn today is to use them, a lot".
As for "what are they really capable of", I can't give a clear answer. They do make easy stuff easier, especially outside of your comfort zone, and seem to make hard stuff come up more often and earlier (I think because you do stuff outside your comfort zone/core experience zone ; or because you know have to think more carefully about design over a shorter period of time than before with less direct experience with the code, kind of like in Steve Yegge's case ; or because when hard stuff comes up it's stuff they are less good at handling so that means you can't use them).
The lower bound seems to be "small CLI tool", the higher bound seems to be "language learning app with paid users (sottaku I think? the dev talks on twitter. Lots of domain knowledge in japanese here to check the app itself) ; implementing a model on pytorch by someone that didn't know how to code before (00000005 seconds or something like this on twitter, has used all these models and tools a lot); reporting security issues that were missed in cURL", middle bound "very experienced dev shipping a feature faster and while doing other things on a semi mature codebase (Ghostty)", middle bound too is "useful code reviews". That's about the best I can give you I think.
- "cop" as an abbreviation for "cross-origin protection" is delightful
- Interesting to see how programmers seems to be separating into people embracing those tools and people rejecting them. I wonder if it's linked to liking the act of coding itself vs liking the results.
- Pouring grease on a fire will make it worse. Picking up a bear cub when the mother is nearby will increase (by how much I don't know) the risks of getting attacked by a bear. Both of those sentences to me sound like genuine description of the reality we live in. Climate change is real and caused by human emissions is another one of those, though a bit less precise as "climate change" is less precise. Saying we should or shouldn't do something is something different.
Also, you can increase power capacity by a lot while reducing emissions, with stuff like solar panels or nuclear power.
- I don't think you should consider this as "winning the lottery", the author has been using these tools for a while.
The sibling comment with the writeup by the creator of Ghostty shows stuff in more detail and has a few cases of the agent breaking, though it also involves more "coding by hand".
- I've known lots of people that don't know how to properly use Google, and Google has been around for decades. "You're using it wrong" is partially true, I'd say more something like "it is a new tool that changes very quickly, you have to invest a lot of time to learn how to properly use it, most people using it well have been using it a lot over the last two years, you won't catch up in an afternoon. Even after all that time, it may not be the best tool for every job" (proof on the last point being Karpathy saying he wrote nanochat mostly by hand).
It is getting easier and easier to get good results out of them, partially by the models themselves improving, partially by the scaffolding.
> non-developer public is mostly convinced that AI is actually artificial intelligence, rather than a very sophisticated next-word predictor
This is a false dichotomy that assumes we know way more about intelligence than we actually do, and also assumes than what you need to ship lots of high quality software is "intelligence".
>While claiming that an LLM cannot follow a simple instruction sounds, at best, very unlikely, it remains true that these models cannot reliably deliver complex work.
"reliably" is doing a lot of work here. If it means "without human guidance" it is true (for now), if it means "without scaffolding" it is true (also for now), if it means "at all" it is not true, if it means it can't increase dev productivity so that they ship more at the same level of quality, assuming a learning period, it is not true.
I think those conversations would benefit a lot from being more precise and more focused, but I also realize that it's hard to do so because people have vastly different needs, levels of experience, expectations ; there are lots of tools, some similar, some completely different, etc.
To answer your question directly, ie “Why do LLM experiences vary so much among developers?”: because "developer" is a very very very wide category already (MISRA C on a car, web frontend, infra automation, medical software, industry automation are all "developers"), with lots of different domains (both "business domains" as in finance, marketing, education and technical domains like networking, web, mobile, databases, etc), filled with people with very different life paths, very different ways of working, very different knowledge of AIs, very different requirements (some employers forbid everything except a few tools), very different tools that have to be used differently.