Ryzal Kamis

trying to build stuff. AI engineer by day, whatever my son needs me to be by night.

[dax @thdxr]

please shut the fuck up i don't even care about the specific thing you're saying i'm just so tired of hearing predictions one after the other telling me what the future is going to be like just please shut the fuck up

[ThePrimeagen @ThePrimeagen]

no amount of vibe coding is going to prevent you from learning how to program

[Adrian Dittmann @AdrianDittmann]

It's just a lot of pressure ...

[Joel Stransky ✨ @JoelStransky]

I hate me

[Hamel Husain @HamelHusain]

LLM bullshit knife, to cut through bs

RAG -> Provide relevant context
Agentic -> Function calls that work
CoT -> Prompt model to think/plan
FewShot -> Add examples
PromptEng -> Someone w/good written comm skills.
Prompt Optimizer -> For loop to find best examples.

[Ryzal Kamis @ryzalkamis]

@benjamincrozat @levelsio @czue Those whose content are stolen, used as training data with no compensation or at the very least attribution?

[kitze @thekitze]

this is the future

[Yann LeCun @ylecun]

I want a laptop with a tensor processor and a unified memory system that runs Linux.

Basically, something that can run large models without weighing a ton, requiring a separate GPU RAM, draining the battery in minutes, costing a fortune, and running a proprietary OS.

[Teknium 🪽 @Teknium]

Why do almost no papers release code, datasets, info on replication, final models, or any combination of these? I thought for science to work results had to be reproducible and verified. Really not scientific and I don't know why academia accepts this

[Ryan Florence @ryanflorence]

I’m tired of native apps.

> download this app to register your kid for tennis

> download this other app to see your the tennis schedule

> download this app to pay your bill

> download our app
> download our app!
> download our app!!

Websites are great, btw

[✨ Jean Yang ✨ @jeanqasaur]

In software engineering, there's what people think of as "docs" and then there's "the information I needed to debug that thing that took to days of my life."

The second thing is different and should be called something other than "documentation." But what should we call it?

[Shit User Story @ShitUserStory]

As a…
– Twitter user
I want…
– unlimited interactivity – centered in audio, video, messaging, payments/banking – creating a global marketplace for ideas, goods, services, and opportunities.
so that…
– It’s no longer possible to even parody this fucking shitshow.

[Sanyam Bhutani @bhutanisanyam1]

Easily the best paper on current State of LLMs! 🙏

A 50 page read but it’s not “just another” survey paper, that only documents facts. The authors actually add very useful commentary capturing all aspects of building Large Language Models.

Hence the result is a collection of ideas we might have missed across months of research.

It covers both building LLMs and effectively applying them to domains, with a focus on current limitations and “sharp edges”

As always, I think great content makes you discover missing bits in your knowledge, for this reason it’s a solid cover to cover read recommendation:

https://t.co/DjCSS8o2Tp

[ThePrimeagen @ThePrimeagen]

I cannot stop thinking about this. Andrew created Zig and ~15 years ago asked on SO how to open a file in C++

[Ryzal Kamis @ryzalkamis]

@jeremyphoward @chrisalbon @neilbhav With mounts, why not? Performance issues?

[Jeremy Howard @jeremyphoward]

The most powerful and compelling writing I've seen on AI safety was a random HN comment from some anon poster.

knodi123, whoever you are, thank you...

[Feross @feross]

I wish more developers understood the constant stream of malware that is posted to npm, PyPI, and all package managers...

Here's just a taste of some crazy malware Socket identified in the past couple weeks...

All malware descriptions were FULLY WRITTEN by Socket AI.

[Shit User Story @ShitUserStory]

I am ready.

#WWDC23 #VisionPro #Apple

[ThePrimeagen @ThePrimeagen]

the thing about using ChatGPT for dev is that i am building tools that require TONS of context on already built APIs and Data Types

everything i see for ChatGPT is building "something new." Not integrating it with a workflow in an area with TONS of tribal knowledge