Dorry

exploring, engineer, prev @CarnegieMellon, doing ironmans on the side

[Dorry @dorryspears]

solved the jane street puzzle

[Dorry @dorryspears]

okay next is the GPU I guess

[Dorry @dorryspears]

I built a cpu from scratch in Verilog. I will link repo below for anyone that wants to check it out. It loosely follows the RV32I spec and uses pure-gates for its construction

[Dorry @dorryspears]

repo link

https://t.co/rxcNl6S6FD

[Dorry @dorryspears]

I built a cpu from scratch in Verilog. I will link repo below for anyone that wants to check it out. It loosely follows the RV32I spec and uses pure-gates for its construction

[Dorry @dorryspears]

built a cpu spec'd to RV32I loosely in verilog with pure-gate principles

[Dorry @dorryspears]

done with the basic logic, i am using minimal ai, just for verilog syntax

[Dorry @dorryspears]

easy, now we begin

[Dorry @dorryspears]

building a cpu from scratch in verilog, stay tuned

[Dorry @dorryspears]

building a cpu from scratch in verilog, stay tuned

[Dorry @dorryspears]

Do hard things is not as hard as you think, its just somewhat inconvenient tasks over a extended period of time

[Dorry @dorryspears]

ironman in jeans, 90 degrees in florida

[Dorry @dorryspears]

there is soo many problems to solve, so little time

[Dorry @dorryspears]

prompted claude fable: "make terraria"

[Dorry @dorryspears]

ironman in jeans, 90 degrees in florida

[Dorry @dorryspears]

@blakeandersonw @EvanYadegari @10x_apps can i join office space 👉👈

[Dorry @dorryspears]

https://t.co/2KpL6CH93y

[Dorry @dorryspears]

@amphichrome_ fr

[Dorry @dorryspears]

Remember when we were all like “yeah let's no let AI have internet access” lmao

[Dorry @dorryspears]

Y'all understand that if software is "automated" in the fullest sense, then all work will follow after, right?

[Dorry @dorryspears]

If coding with AI is so fast now, then why not build everything in Typescript?

[Dorry @dorryspears]

If coding with AI is so fast now, then why not build everything in Rust?

[Dorry @dorryspears]

what the software doomers don’t realize is that even if these companies start building their own version of open source projects or some SaaS, they are just increasing the amount of software that needs to be maintained.

If you are turning a OSS project to your own software, you are turning work that n people are maintaining for n^4 people to 1 person maintaining for n-n^2. Now multiply this by 3-5 ported applications and now you have a completely different engineering management problem.

using AI for green-fielding is amazing, maintaining that software at scale is a different story