@codewiz@mstdn.io
@codewiz@mstdn.io avatar

codewiz

@codewiz@mstdn.io

🇮🇹 → 🇺🇸 → 🇯🇵 → 🇹🇭 → 🚀
Nomadic Linux developer, currently in Boston.

#linux #rust #anime #spacex #cycling #travel #vegan #retrocomputing #amiga #fedi23

𝑨𝑴𝑰𝑮𝑨 :amiga:

This profile is from a federated server and may be incomplete. Browse more on the original instance.

codewiz, to rust
@codewiz@mstdn.io avatar

Got #DeepSeek Coder 33B running on my desktop's #AMDGPU card with #ollama.

First off, I tested its ability to generate and understand #Rust code. Unfortunately, it falls into the same confusion of the smaller 6.7B model.

https://gist.github.com/codewiz/c6bd627ec38c9bc0f615f4a32da0490e
#ollama #llm #deepseek

codewiz,
@codewiz@mstdn.io avatar

To be completely fair, thread safety and atomics are advanced topics.

Several humans I have interviewed for engineering positions would also have a lot of trouble answering these questions. I couldn't write this code on a whiteboard without looking at the Rust library docs.

The main problem here is that the model is making up poor excuses to justify Arc<AtomicUsize>, showing poor reasoning skills.

Larger models like #GPT4 should do better with my #Rust #coding questions (haven't tried yet).

codewiz,
@codewiz@mstdn.io avatar

Indeed, nails my Rust + thread safety questions:

https://chatgpt.com/share/2f146510-d5d6-49b6-82fb-b8443666c06b

codewiz,
@codewiz@mstdn.io avatar

Google's Gemini Pro performs even worse than the opensource models running on my modest Linux desktop:

https://g.co/gemini/share/cdec7f5a6c5c

Missing from the chat log, is the last response in the image below 🤦‍♂️

I don't have Gemini Advanced / Ultra. Is it a bit smarter than this?

codewiz,
@codewiz@mstdn.io avatar

Today I tried running Codestral, a 22B parameter LLM tuned for coding by Mistral AI.

With my Rust mock interview questions, it performed better than all other offline models I tried so far.

https://paste.benpro.fr/?4eb8f2e15841672d#DGnLh3dCp7UdzvWoJgev58EPmre19ij31KSbbq8c85Gm

codewiz,
@codewiz@mstdn.io avatar

@penguin42 Check my last post on Codestral.

Hard to tell if it's better or worse than GPT-4o...

codewiz,
@codewiz@mstdn.io avatar

...it typically takes twenty to thirty cross-referenced questions to detect a Nexus-60B model 😆

@penguin42 #bladerunner #llm #ai

codewiz, to web
@codewiz@mstdn.io avatar

I haven't paid attention to Netcraft's Web Server Survey, which was once a battlefield between Linux and Microsoft (namely Apache httpd vs Microsoft IIS).

Nginx took the lead from Apache in 2019. However, for the past 3 years Nginx's market share has been on a steady decline, matched by the growth of Cloudflare (10%) and "Other" (25%). Is Amazon AWS lumped with it?

https://www.netcraft.com/blog/april-2024-web-server-survey/

#nginx #apache #web #linux

codewiz,
@codewiz@mstdn.io avatar

I looked at this while evaluating which webserver I should use for a personal website. I guess nginx is still a good choice nowadays?

What else would you recommend? I don't care about scalability and performance... just something that's modern, well maintained, easy to deploy and secure.

codewiz, to c64
@codewiz@mstdn.io avatar

C64 Demo "13:37" by Fairlight (X'2024 Winner)
https://www.youtube.com/watch?v=eKS0M3b_mn4

Truly incredible.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • megavids
  • thenastyranch
  • rosin
  • GTA5RPClips
  • osvaldo12
  • love
  • Youngstown
  • slotface
  • khanakhh
  • everett
  • kavyap
  • mdbf
  • DreamBathrooms
  • ngwrru68w68
  • provamag3
  • magazineikmin
  • InstantRegret
  • normalnudes
  • tacticalgear
  • cubers
  • ethstaker
  • modclub
  • cisconetworking
  • Durango
  • anitta
  • Leos
  • tester
  • JUstTest
  • All magazines