faragon

Why Microsoft just don't make optional the TPM requirement for upgrading from Windows 10 to 11?

De-centralized computing is similar to what happened in IBM/Sony Cell CPU: independent SRAMs with an associated vector unit, doing transfers to the main/host DRAM via DMA using a high performance ring buffer. The benefit is simpler architecture, but more difficult software implementation, although for AI or graphics development behind an abstraction could be perfectly fine.

That should be a property of a web site. E.g. the web site using a property in the Cloudflare configuration. That way there would be competition between websites allowing or not IA agents on user accounts.

I would like to know the projected global impact in productivity increase because of that investment.

The more YouTube does that, the more I use Spotify.

The specs:

https://amiga-news.de/en/news/comments/329201.html

Unreal engine.

Yes. And that can happen with and without IA.

Translated: same engineers are more productive by using IA.

Also in Spanish, "viernes santo" ("holy Friday")

The main use I find for LLMs is code review and corrections following a list of criteria. It helps to detect overlooked issues.

It is also useful for learning from independent code snippets, for e.g., learning a new API.

If a person can read copyrighted material and produce derivative works, why not an AI?

The AI assistant for Office 365 is confusing, e.g., the "light bulb" feature in meetings/calendars adds random OneDrive references. This can be dangerous, as it might be mistaken for attached documents, leading to misunderstandings between the people in the meeting. Such features should be OFF by default and enabled selectively, allowing users granular control over their impact.

LZ77 compression/decompression in pure Bash using hashmaps:

https://github.com/faragon/lzb/blob/master/lzb

I hope Google puts also the same effort in tuning the models. E.g. the difference between Claude/ChatGPT and Gemini is abysmal (not on synthetic scores, but how it feels).

Improving audio with AI would be a good thing for @Google/@YouTube to put time/effort in! :-)

Working with QNX today is like working with 25 year old software: tiny community, few examples, lots of issues for which you end doing avoidance workarounds, etc.

I would have more interest on it if being open source.

Very beautiful implementation of the awk interpreter in less than 600 lines!

https://github.com/Gandalf-/coreutils/blob/master/Coreutils/...

What I find incredible is that Google has had the knowledge and resources to do this for at least five years, yet they're still milking the "old cow". It reminds me of Intel sitting on their money while a near-bankrupt AMD sped past them.

I was expecting here SDR radio processing: noise read from GPIO's passed to a web browser with a pipe, showing activity on all channels at once.

Before reading the article I was expecting using 1-bit instead of bfloats, and using logical operators instead of arithmetic.

I used Cygwin for more than a decade. I prefer Git Bash (msys-based).

I'm glad Microsoft uses Bash in the example, instead of their own Windows shells. As a user I would like having something like "Git Bash" for Windows built in the system, as default shell.

Thank you

Using indexes instead of pointers in data structures works well, and the cost of the base address + offset is negligible, as similar address calculation is already generated by the compiler when accessing into an element of a data structure. In addition to that, mention that indexes can be used as offsets, or as an actual indexes of the size of an individual element, i.e. in that case non-trivial data structures with e.g. >= 32-byte elements could address hundreds of gigabytes of RAM.

A practical use could be e.g. using bit fields can be convenient, e.g. having 32-bit indexes, with the higher bit for the color in a Red-black tree. And in case of requiring dynamic-sized items in the tree nodes, these could be in different 32-bit addressable memory pools.

I hope cases and gender get simplified as well. German with those simplifications would be much easier to learn.

Slower, clock wise. My guess, considering the IPC, is that it could perform better.

People could prompt for authenticity, adding subtle mistakes, etc. I hope that AI as a whole will help people writing better, if reading back the text. It is a bit like "The Substance" movie: a "better" version of ourselves.

This reminds me Henri Poincaré's "Science et méthode" (1908)