• 0 Posts
  • 5 Comments
Joined 11 months ago
cake
Cake day: August 21st, 2023

help-circle

  • Yeah, it’s technically complicated. The “5” in RISC V kinda denotes the history involved in this (IIRC). I’ve seen a bunch of stuff on this, but it has been awhile and my attempt to focus may misconstrue the blur in my mind. So don’t take this as gospel. The previous attempts at a completely open ISA were failures in that they weren’t compatible with the foundry toolchains or peripheral business licenses and tech. They tried to do too much or force change tenured elements that have no open source replacement. RISC V was the first to be practical and garner support from industry and academia. There was a major hurtle that it had to overcome early on. As soon as this happened, the sale of ARM was announced shortly thereafter. I forget the details. It was in a conference talk on YT IIRC. I just recall thinking, that the timing of the sale of ARM spoke volumes about what RISC V is in the grand scheme of things. In 10 years time, I think everything will be RISC V. It will dominate in every sector from small microcontrollers all the way to the largest data centers and everything in between. It will be the final blow to x86, ARM, and anyone that fails to migrate quickly. This is Linux for hardware and fixes a lot of problems that are unaddressed with this proprietary crap designed in closed rooms by a few hundred people. No company or group is smarter than the whole world with unrestricted access.


  • RISC V is the first open source Instruction Set Architecture. Everything needed to fab the chip is open source. It is like how Linux dominates the world in enterprise software, RISC V is doing that for chip fabs. ARM is a closed ISA. Making the chips requires paying a royalty to ARM for each chip made. It is a scheme to extract money through manipulation not true competitive innovation or value.

    This is like any other market, as soon as an equivalent open source alternative exists, the incumbent extortionist’s days are numbered. Its only customer base is from convenience or ignorance.



  • I’ve written several python scripts to extend the capabilities of a text to speech AI using an offline Llama2 70B model. It takes enthusiast level hardware to run. The few errors it makes in code snippets it can also correct by pasting the terminal error message and prompting.

    I don’t write python, haven’t had to go online to look up anything about Python syntax, and have a script that can take any text and convert it to speech, convert any wave file into text, or concatenate any wave files.

    It takes larger models and some testing to find a good combination for a task. It also helps to use a model that is not subject to public spotlight and political pressure like GPT or Bard. The tiny models like 7B’s and 13B’s are like talking to children or teenagers.