Yeah, I’ve been seeing the same. Purely economically it doesn’t make sense with junior developers any more. AI is faster, cheaper and usually writes better code too.
The problem is that you need junior developers working and getting experience, otherwise you won’t get senior developers. I really wonder how development as a profession will be in 10 years
Working on a big codebase, I don’t even get the idea to ask an AI, you just can’t feed enough context to the AI that it’s really able to generate meaningful code…
That’s not a hard limit, for example google’s models can handle 2-million-token context window.
AI isn’t ready to replace programmers, engineers or IT admins yet.
On the other hand… it’s been about 2.5 years since chatgpt came out, and it’s gone from you being lucky it could write a few python lines without errors to being able to one shot a mobile phone level complexity game, even with self hosted models.
Who knows where it’ll be in a few years
Since I already use ZFS for my data storage, I just created a private dataset for sensitive data. I also have my services split based on if it’s sensitive or not, so the non sensitive stuff comes up automatically and the sensitive stuff waits for me to log in and unlock the dataset.
Svn: 20 October 2000
Git: 7 April 2005
I remember using svn when git development was started
which is about exactly as old as git.
Wdym by that?
And then it doesn’t work on one VM and it takes you three days to figure out that VM had a slightly different version of a library and that makes the app segfault.
Or a few more gb of LLM?
No, all sizes of llama 3.1 should be able to handle the same size context. The difference would be in the “smarts” of the model. Bigger models are better at reading between the lines and higher level understanding and reasoning.
Wow, that’s an old model. Great that it works for you, but have you tried some more modern ones? They’re generally considered a lot more capable at the same size
Increase context length, probably enable flash attention in ollama too. Llama3.1 support up to 128k context length, for example. That’s in tokens and a token is on average a bit under 4 letters.
Note that higher context length requires more ram and it’s slower, so you ideally want to find a sweet spot for your use and hardware. Flash attention makes this more efficient
Oh, and the model needs to have been trained at larger contexts, otherwise it tends to handle it poorly. So you should check what max length the model you want to use was trained to handle
I still use http a lot for internal stuff running in my own network. There’s no spying there… I hope … And ssl for local network only services is a total pita.
So I really hope browsers won’t adapt https only
But even if you use GoMommy extra super duper triple snake oil security checked ssl cert, if I trick LetsEncrypt to sign a key for that domain I still have a valid cert for your site.
I doubt the disk will bottleneck at 40mb/s when doing sequential write. Torrent downloads are usually heavy random writes, which is the worst you can do to a HDD.
I was trying to find an article I read about a year ago, about an experiment where AI was assisting a doctor. Where it suggested questions and possible diagnosis for the doctor to look into.
IIRC the result was both faster and more accurate diagnosis. Too bad I can’t find it again now :(
You’re not great taking medical advice from a doctor either, seeing how often they’re wrong.
I remember back in the day this automated downloader program… the links had a limit of one download at a time and you had to solve a captcha to start each download.
So the downloader had built in “solve other’s captcha” system, where you could build up credit.
So when you had say 20 links to download you spent some minutes solving other’s captchas and get some credit, then the program would use that crowdsourcing to solve yours as they popped up.
Llama3 8b can be run at 6gb vram, and it’s fairly competent. Gemma has a 9b I think, which would also be worth looking into.
And just to top it off, make this pythonscript a dialect of rust
Well, it wasn’t a comment on the quality of the model, just that the context limitation has already been largely overcome by one company, and others will probably follow (and improve on it further) over time. Especially as “AI Coding” gets more marketable.
That said, was this the new gemini 2.5 pro you tried, or the old one? I haven’t tried the new model myself, but I’ve heard good things about it.