SPascareli13

That's the smallest dan luu post I've seen.

There's a new SOTA model every few months, are you supposed to buy a new chip every new release?

My thought was that ASICs turned out not to be worth it for crypto mining because consumer hardware evolved fast enough to do it, while also being cheaper and having some resale value, while ASICs are useless besides mining and have no resale value.

So I'm extrapolating this same idea to LLM inference.

Yet it doesn't seem that ASICs will have any particular advantage over consumer hardware since AI is very memory heavy, which is (right now) expensive no matter how you package it. And the compute is just simple matrix multiplication, which is almost entirely what GPUs were meant to do anyway.

Like how crypto used ASICS but then didn't because the scaling of consumer hardware made it obsolete?

But that assumes I have some switch case somewhere right? If I passed an array of "orders" to a downstream function, it knows that what it has is orders, not something else, so it doesn't need to check for anything, and in the case I checked for errors upstream (when I try to allocate a new order) all downstream functions know that no invalid order can be passed, in which case you only have one check at creation time.

That's why I haven't fully understood yet how working like this is simpler.

I never fully understood how to work with this "reserved" values being valid instead of errors when you can't allocate more memory, in either case you still need to check if you have a real entity or a reserved/error, right? Does it really make things simpler?

Amazing writing, very accessible too.

So batching requests is always something I think should increase performance by a lot, but most server implementations make this pretty difficult, but the thing I struggle the most to understand is how to keep the latency down if you have multiple clients request all batched together? The total amount of latency for all clients is always the latency for the slowest.

You need some JS in the middle but yes, you write your image to a buffer and share it with JS, then in the JS you put it in the canvas.

Just trying to learn C again, making things from scratch in a multiplatform way, interfacing with X11 on Linux and wasm on the browser.

It's been fun dealing with memory and C's weird design in this age of agentic coding.

Interesting aside, BYD (Chinese EV automaker) has took over production in the old Ford factory in Camaçari.

If the model is trained to be a interpreter, then that means that the loss should reach 0 for it to be fully trained?

Also, if it's execution is purely deterministic, you probably don't need non linearity in the layers, right?

My current stance with reviewing code is: It's not ok to make another human review the code you made with AI, if you used AI then you're the reviewer, so unless you come to me with a well defined question or decision to make, just merge it and take responsibility.

Obviously that could only work in a high trust environment, that why open source suffers so much with AI submissions.

Freeciv was what brought me too the civ world, I'm sure this project will be the same for many children of this generation.

I really dislike this idea of testing in go: only ever use an interface, never the real implementation + mockgen the mocks based on this interface + use the mocks to assert that a function is called, with exactly this parameters and in this exact order.

I find this types of tests incredibly coupled with the implementation, since any chance require you to chance your interfaces + mocks + tests, also very brittle and many times it ends up not even testing the thing that actually matters.

I try to make integration test whenever possible now, even if they are costly I find that the flexibility of being able to change my implementation and not break a thousand tests for no reason much better to work with.

Gemini is quite optimistic here thinking GTA VI will be released by 2035.

The last batch of juniors we hired just completed 4 years in the company, which I would say is a pretty successful batch, but sadly we haven't hired juniors since.

Edit: I must qualify that this is for software developers only, we did hired juniors for things like data engineers, security, IT and such.

Implement the simplest thing that works, maybe even by hand at first, instead of adding the tool that does "the whole thing" when you don't need "the whole thing".

Eventually you might start adding more things to it because of needs you haven't anticipated, do it.

If you find yourself building the tool that does "the whole thing" but worse, then now you know that you could actually use the tool that does "the whole thing".

Did you waste time not using the tool right from the start? That's almost a filosofical question, now you know what you need, you had the chance to avoid it if it turned out you didn't, and maybe 9 times out of 10 you will be right.

As far as I know there is no way to do Promise like async in go, you HAVE to create a goroutine for each concurrent async task. If this is really the case then I believe the submition is valid.

But I do think that spawning a goroutine just to do a non-blocking task and get its return is kinda wasteful.

Go is a great tool for simulating network concepts, I've done it too.

I think I tested very casually some time ago with Go maps and up to like one hundred items the linear search on array was faster than map lookup. Considering that many times when we use Maps for convenience they will have less than a hundred items this could be useful.

Unfortunately I don't have the results (or the test code) anymore, but it shouldn't be hard to do again (casually at least).

Not sure about this particular instance. LTT got a video taken down by community guidelines violation just the other day, and they are much bigger than Jeff.

> Isn't Jai still vapourware after a decade ?

Yes, I guess. I wouldn't be surprised to see "Jai inspired" languages coming out before Jai itself, since some of the ideas look pretty good.

> How does one even get the compiler and build programs ?

Don't know how seriously you're asking this, but I believe you can apply to use it and if Jon likes your credentials and what you want to try it on he might let you have the compiler for some version. Don't know why he won't just open source it and say it's a early version passive of change, but game development is not usually very open source friendly compared to web dev.

I don't know if there's ever been a language that was announced 10 years before its first closed beta. I don't even know if having a "closed beta" for a language is something that really happens.

So there's a lot that is different with Jai, it's more like a highly anticipated game than a language.

I was thinking about that just the other day, who it would be really cool if Go had compile time code execution. I think Jai is making that a very prominent feature of the language.

This looks absolutely amazing and absurd at the same time.

All IO is synchronous? Writes don't actually write when they say they do? Running 201 queries is as "fine" as running a single one with join?

It seems as they got every bad ideia in the book and somehow made it into something that works (that is, if it actually does).

Only 3.4k of C code for a full http and https server? I honestly thought you would need a lot more for it to be fully compliant with the spec.

I remember a few years ago a very good engineer in our company shared a very similar article about hexagonal architecture in golang, and a lot of people started using it, including myself for some time.

Now he's not here anymore, and posts in his linkedin about simplicity and how people overcomplicate things in software so much. This shows me that you should be careful when taking advice from other engineers, they learn and move on from what was previously a "best practice", while you might get stuck thinking it's worthy it because "that one very good engineer said it was how it should be done".

You need a separate package to actually serialize protobuf from a congen'd struct, so it uses reflection.

Someone said that the word "delve" is a favourite of AI and a sign that something was AI written.