I’m old enough to remember and toy with the now long-dead XNA. It was lots of fun, and gave a lot of us students versed with C# a sort of first-hand exposure with the .NET. If only (the old) Microsoft wasn’t so stupid, short-sighted, and selfish at the time.
If only it were that rosy. I tested a few of the top open-source coding models on a beefy GPU machine, and they all behaved like anything about anything - simply rotating in circles and wasting electricity.
Has anyone used this in earnest with something like OpenCode? Over the past few months I’ve tested a dozen models that were claimed to be nearly as good Claude Code or Codex, but the overall experience when using them with OpenCode was close to abysmal. Not even a single one was able to do a decent code editing job on a real-world codebase.
With M2, yes - I’ve used it in Claude Code (e.g. native tool calling), Roo/Cline (e.g. custom tool parsing), etc. It’s quite good and for some time the best model to self-host. At 4bit it can fit on 2x RTX 6000 Pro (e.g. ~200GB VRAM) with about 400k context at fp8 kv cache. It’s very fast due to low active params, stable at long context, quite capable in any agent harness (its training specialty). M2.1 should be a good bump beyond M2, which was undertrained relative to even much smaller models.
Surprisingly, to me, it’s the other ways around - and, I’ve been writing code for two decades now. I love programming and even with AI, I will always have the last word, but I also realized along the way that programming is only a means to an end - you write code to get something done, not to write the code itself. With AI, I can finally give chance to my hundreds of ideas and see what sticks.
They do. I’ve been teaching cross-career programming courses in the past, where most of my students had day jobs, some, involving hard physical work. They’d gladly swap all that for the opportunity to feed their families by writing code.
Just comes to show how the grass is always greener when you look on the other side.
That said, I also plan to retire up in the mountains soon, rather than keep feeding the machine.
We have a very similar feature on https://feedle.world. Every search has its own dedicated RSS feed that can new followed directly, as well as an iframe that can be embedded on other people’s websites. This way, anyone can build accidental blogrolls, based o topics of interest.
P.S. for people whore not really into RSS, we are also Beta testing the option to subscribe to searches and get results in email digests. Same idea, but you don’t need to bother finding an RSS reader.
reply