tymscar

I think its good especially if you are on an M1 generation. This is one of those releases with loads of fixes and speedups and not a whole lot of anything else

Who promised it?

I find it so strange when someone comments something like this.

Nobody thinks this is okay or that Simon should have to play those games. The point of the reply was that while this is stupid, the solution is trivial.

I’m genuinely very curious what the point of comments like this is. I am not joking, I want to understand.

I see examples like it 50 times a week in random places, and nobody walks me through their mental state to help me understand what they thought the point was.

This is awesome, thanks for sharing

I think its a very fair thing you have flagged!

Well it doesn’t matter because the bottleneck here is actually quite small for me. The issue is vram. If anything the bottleneck is my 4080.

Thanks!

I did get feedback on all sorts of things over the years.

One of them was to do with sentence lengths.

I would bet my trillion-dollar company on it because I understand how RAM works.

If you use NixOS, you don’t need to do any of that anymore. A single command in the CLI, and you have everything I do.

I did mention that in my case I had the 4080, but you do not need the 4080 whatsoever. You can run the same model with less context on a single £200 V100 or with the same amount of context on two of these.

Well I did in fact meaningfully combined them without an issue, that was the whole point of the blogpost.

Building computers is very easy. I would suggest watching a YouTube video to get the general gist, and then once you buy the parts, just Google for whatever doesn’t go well.

I built my first Pentium 4 one when I was like six, so I’m sure someone much older that’s into tech can do it without an issue.

There are also tons of Discord communities that are willing to help you live if you encounter any issues.

GPTzero is a joke.

It thinks this is AI: “I bought a datacenter GPU that doesn’t even have a normal PCIe connector, stuck it in my gaming PC with an adapter, and now I have 32GB of VRAM across two GPUs running a 27 billion parameter model at 32 tokens per second.”

There’s nothing AI about that. Not all SOTA LLMs agree, hell, none of them do. The same exact example I sent here gives me 0% in some, 10% in others, 100% in GPTzero.

Neither does this one. I replied in another thread. It comes out as 0% and the one from 2021 comes out as 8%. LLM detectors are all BS

I have written the blog post. I know empirically that I have used 0% AI while writing it. I also know LLM detectors are total BS and they don't really work. I have tried a couple on this exact blog post, and QuillBot, for example, gave me 0% AI detected on it.

I have then used a blog post of mine from 2021. QuillBot gave me 8%...

The King James version of the Bible came out at almost 100% AI generated a while ago. It was the HN front page.

Stop thinking that if someone writes in a way that is fun or looks like what you would think an AI writes, then it is AI generated. Loads of the time it is, but sometimes it's not, and it really hurts those like me.

I’m glad I’ve started this blog before the AI wave so I can prove people I’m just weird at writing.

It grinds my gears how so many people just talk about my writing style instead of the content.

FYI, not a single line was AI written. If there is a hallucination, it’s fully mushy brain sourced.

Agree. But I have not used AI in the slightest.

Some of us just write that. AIs had to learn it from somewhere.

AI didnt edit a single word of this post.

Not at all, no. I had this chat before about how I am one of those unlucky few that loved the way LLMs write nowadays since the mid-2000s.

Slowly but surely, I had to remove my beloved lists, emojis (though LLMs do less of that now, maybe I can incorporate them back), and emdashes.

But with miles better support, thats why I went this route. Cuda is hard to beat

Thats why I did it. I think it’s important to put things like that into perspective

So it didnt end up working too well seeing that the latest data is from 4 years ago.

Jerryrigseverything has a video of him opening one up. Not really what you asked for but it scratches the itch

I still remember the massive shift to SDRP and HDRP. Honestly, now in retrospect, almost a decade later, I think it was clearly done wrong. It was a mess, and switching over was a multi-week procedure for anything more than a hello world program, and what you got in return wasn’t something that looked better, just something that had the potential to.

Similar story with the whole networking stack. I haven’t used Unity in years now after it being my main work environment for years, but the sour taste it left in my mouth by moving everything that worked in the engine into plugins that barely worked will forever remain there.

Im sure its partly skill issue

I think you might be answering to a different person or misunderstanding what I said but you are right that just as I don’t have an intuition for where the acuity shows up in the corpus, I don’t think Claude does either

Thats cool.

I do wonder where that extra acuity you get from 1% more shows up in practice. I hate how I have basically no way to intuitively tell that because of how much of a black box the system is

Or you could just have a cli that does that, no MCP needed

No? You can just have env vars

So basically the best way to use MCP is not to use it at all and just call the APIs directly or through a CLI. If those dont exist then wrapping the MCP into a CLI is the second best thing.

Makes you wonder whats the point of MCP