mhast

I recall seeing a audio technician test some of the modern VSTs and it seemed extremely impressive to me as an amateur.

It's to the point where you virtually put microphones around the virtual instrument and it behaves (according to the person who does it for a living) creates an intuitively correct sound.

Edit: and to be clear this was a demo from before the LLM times so it's something built the traditional way.

IIRC you can get the download link if you go to view "your content" on the Kindle website.

And if a book is on the no DRM list you can find a "download" link in the drop down menu for the book.

But I have not found a way to determine if a book has this option before buying it. So I'm not longer buying anything on Amazon.

There was a new jailbreak released quite recently which was quite painless to go through. IIRC it's called Sanctuary.

The only problem is if your device is connected to wifi and has updated to new firmware.

Chris from "Abroad in Japan" has also done a couple of Japan biking trips. He also has quite a few travel logs from less obvious places in Japan. (Including some very obvious ones, like Mt Fuji.)

They are edited to be a bit more "dramatic" but once you get into the actual day to day stuff they are quite chill.

https://www.youtube.com/watch?v=zTHXWPeJnNA

Almost 30 years ago there was a project about making a computer which used FPGAs for all "software". It was called RAW.

Baring it all to software: Raw machines | IEEE Journals & Magazine | IEEE Xplore https://share.google/nI9GyFJu4HvbYFrBf

(If you Google the name you will find free PDFs as well, the IEEE page is more useful as a summary and such.)

In the long term interconnected nations might make electricity cheap everywhere. Currently it's mostly making it more expensive everywhere else. (As someone living in south of Sweden.)

Edit: But to be fair... The biggest reason it's expensive is likely that companies like to make money.

For the problems ASICs exist they vastly outperform general hardware. Typically both in absolute speed and efficiency.

But it's only possible to make custom ASICs when you have a specific problem to solve. For newer crypto systems they can vary enough parameters that building a flexible enough ASIC to recoup the investment before the algorithm changes and makes your hardware useless.

For problems where the problem to solve remain in the problem space the ASIC can solve there is no point to use a thing else.

The different models of chips and memory config have very different memory speeds as well.

Eg the M4 Max 128GB has a bandwidth speed of 500GB/s+. And that's true for other models as well.

But as you note, the base speed has also increased over the versions.

The main PCB is pretty much everything that has value. The rest is a heatsink, case and PSU.

The PC platforms have anemic memory bandwidth in comparison. Eg, Strix Halo is 256GB/s max. If money is a bigger limiter than performance it can be an option though. As can Nvidia DGX Spark machines. (Also limited to 128GB memory and comparatively low bandwidth, but higher compute than Strix Halo.)

It's worth noting that the 5090 (or the RTX Pro 6000 big brother with 92GB VRAM) will run rings around the Mac when it comes to compute.

My old 3090 is typically significantly faster (almost 2x token/s) than my M4 Max 128GB machine, as long as the model fits in the 24GB of VRAM.

In most situations it's a better idea to just buy tokens. But there are definitely cases when that's not an option. And then a machine like the M5 Ultra can allow you to do things locally for a fairly limited budget. And in a simpler package to manage than a machine with multiple GPUs.

One case where Apple can be very useful is if you're traveling a lot for work.

Wherever you are you can probably find a local store that sells and services Macbooks. So if your machine is damaged during a trip you can get it fixed, or buy a replacement on the spot, and at least get through the work you are there to do.

For small-ish companies that can be a very useful feature.

I think the point was that even if you managed to make a O(n*2) parser it will ve fast enough for human entered problems.

Sounds like this is similar to a decompilation (pay walled article so I haven't read it).

A while back a listened to a podcast where they interviewed people doing this and it was quite interesting. (And it's an engineering podcast so many technical details.)

https://softwareengineeringdaily.com/2024/05/24/bonus-episod...

Their Foundation adaptation is pretty terrible at understanding what the books are about.

Going from a story where societies and not individuals are the "heroes" to people with magical powers.

I'm assuming it's meant the other way around.

Given their estimates of a Mac being able to generate $1k (per month?) a 5090 with a lot more power would be able to generate $50k. For a $3k piece of hardware. Which is obviously not realistic. (As in, nobody is paying that much for the images, which seems to match well with no actual requests on the system.)

Strange. I've been running it on several different ubuntu 24 04 machines with standard terminal with no issues.

But with waymos it would be possible. Mark those streets as "extremely slow" and never go there unless you are dropping someone off. (The computer has more patience than human drivers.)

If that's too annoying then bad parking by school areas so the situation doesn't happen.

The field of "AI assisted programming" is progressing extremely rapidly.

Claude code (arguably the most recent large change, even if it wasn't the first of it's type) was released one year ago.

After watching the video I'd say that it is similar to my own reaction when opening my own code that is 2 years old. (To be clear, code I myself wrote 2 years ago, without AI.) Or even more realistically, code I wrote 6 months ago.

Buy it mostly reads like someone making an exaggerating claim to get s boost from a populist narrative.

If you just want to play I believe the Google alternative can even run on the free tokens you get from them. It's not going to do all that much before running out of tokens but you can probably have it make a simple single page web site for a company or something like that.

Even at the Max plan it's not unlimited. IIRC they say the limit is something like 20x the $20 plan.

With "normal use" you're unlikely to hit the limit though, unless you only use the Opus model and do things in parallell.

If you want to fingerprint text you can also do it by small insignificant changes to text which doesn't change the meaning.

If you have a number such locations with alternatives then you can make a number of identifiable versions by combining alternates.

The typical "best practice" for these tools tend to be to ask it something like

"I want you to do feature X. Analyse the code for me and make suggestions how to implement this feature."

Then it will go off and work for a while and typically come back after a bit with some suggestions. Then iterate on those if needed and end with.

"Ok. Now take these decided upon ideas and create a plan for how to implement. And create new tests where appropriate."

Then it will go off and come back with a plan for what to do. And then you send it off with.

"Ok, start implementing."

So sure. You probably can work on this to make it easier to use than with a CLI chat. It would likely be less like an IDE and more like a planning tool you'd use with human colleagues though.

It is typically neither free nor open to develop on consoles. As in, you pay to access developer tools.

iOS and Android less so (even if there is a one time charge for Android and a yearly charge on Apple). OTOH I have not heard of them usually reaching out to more than a handful of devs for promotion purposes.

Jim Keller goes into some detail about what difference the ISA makes in general in this clip https://youtu.be/yTMRGERZrQE?si=u-dEXwxp0MWPQumy

Spoiler, it's not much because most of the actual execution time is spent in a handful of basic OPs.

Branch prediction is where the magic happens today.

That would probably be preferable from a PR standpoint.

None of those games are made for children. Particularly not CoD and Battlefield.

That doesn't mean kids don't want to play them. But Roblox is pitching itself as a safe space for kids.

"People Make Games" on YouTube made some videos about it. Recommended watching for any parent who is trying to figure out if they should allow their child on Roblox. (Spoilers, the answer is "no".)

There are Python bindings for the framework as well.

Personally I could see Go being quite nice to use if you want to deploy something as eg a compiled serverless function.

I'm assuming the framework behaves the same way regardless of language so you could test using Python first if you want and then move over to eg Go if needed.

I have a M4 Max with 128 GB memory. Even on that machine I would not consider 70B+ models to be useable. Once you go below 20 tokens/s it becomes more like having a pen pal than an AI assistant.

MoE models can still be pretty fast. As are smaller models.

(This is mostly a warning for anyone who is enamored by the idea of running these things locally to make sure to test it before you spend a lot of money.)

Currently I'd probably say the Nvidia RTX pro 6000 is a Challenger if you want local models. It "only" has 96 GB of RAM, but it's very fast (1800 GB/s). If you can fit the model on it and it's good enough for your use case then it's probably worth it even at $10k.