Similar User
Marco 🦦🐧 photo

@PwzKB

Paul Bostwick photo

@PaulBostwick15

Carly Brousek photo

@CoachBrousek

Ji Ung Lee photo

@JiUngLee1

Sibelica photo

@estemujeron

Carina Grupillo photo

@CGrupillo

rapid link photo

@rapidlink3

NoBody photo

@nobody_are

Will Hewett photo

@WillHewett

woody photo

@poomwar

hendy 🦊🥍 | 📖 tian guan ci fu photo

@retrohryuck

Quincy Snoddy Jr photo

@_mixtape_1

Carl Kurlander photo

@Ckurlander

Dj Femmy photo

@DjFemy8

Louinjaxxx photo

@Louinjax

P Reposted

Undercover.. 😅


P Reposted

Today, my dad got a phone call no parent ever wants to get. He heard me tell him I was in a serious car accident, injured, and under arrest for a DUI and I needed $30,000 to be bailed out of jail. But it wasn't me. There was no accident. It was an AI scam.


P Reposted

Elephant carefully tests the electric fence before removing it..


P Reposted

Sheep dogs don't get enough credit


P Reposted

Surprised not to see many say this: OpenAI just declared the *end* of a paradigm, more so than the beginning of a new one. Scaling inference is powerful and interesting; same as two years ago. What's new? Declaring that scaling up the model itself & doing standard RLHF is over.

With the small model size of gpt-turbo, the strong reliance on human feedback, and the addition of tools/plugins, OpenAI (like many others) is silently giving up on the notion that scale is all you need, but without conceding that. This is a welcome step, but it should be noted.



P Reposted

Few species have been observed imitating human behavior, but the list of observations has recently grown thanks to social media videos. This is just a deer scoring a goal and celebrating.


P Reposted

It's not that Bayesian epistemology has caused all the trouble in the world, it's that all the trouble in the world has caused Bayesian epistemology. @DavidDeutschOxf


P Reposted

AI, recently for example, GPT and ChatGPT is a wonderful thing and can be very useful and it has nothing whatever to do with AGI. In fact, as I've written, it's more or less the exact opposite of AGI because it involves honing the program to conform more and more precisely and in…


P Reposted

Richard Feynman on how to study ✍️

Tweet Image 1

P Reposted

Orbits. Made with #python, #matplotlib and #numpy.

Tweet Image 1

P Reposted

Consider DSPy's SignatureOptimizer by @michaelryan207 @kristahopsalong It generalizes OPRO to multi-step programs: it proposes instructions based on signatures of each module + previous attempts of the pipeline. If it could see the metric + failures, it'd be more grounded.

Tweet Image 1

P Reposted

have you ever seen a computer glitch out this bad. personally i have not. (this isn't glitch art, it's just incompetence)


P Reposted

🦜LangChain <> DSPy🧩 DSPy is a fantastic framework that **compiles** your LLM program Specifically, the DSPy compiler will internally trace your program and then craft high-quality prompts for large LMs to teach them the steps of your task. Thanks to @lateinteraction we have…

Tweet Image 1
Tweet Image 2

P Reposted

Paris AI is on fire.

The ML ecosystem in France is on fire🔥 It has amazing talent and resources. Here are 10 facts you might not know: 1. There are great research labs - from @MistralAI and @kyutai_labs to large ones from @AIatMeta and @GoogleDeepMind The Llama 2 and CodeLlama authors are based in…

Tweet Image 1


P Reposted

My mom used to take my mouse and keyboard away when I didn’t finish my homework

Today, we unveil the first public demo of the MouthPad^: smart mouthware for hands-free computer control. Witness Corbin swiftly navigate Google Maps on his phone using only subtle tongue movements. Cursor control, clicks, and swipes: full access.



P Reposted

oh no


P Reposted

people in 20 years with AI

Tweet Image 1

P Reposted

A common response to many AI efforts is "but humans have that problem too". I don't know who is to blame here (maybe the AGI nonsense?) but most of us are not trying to replicate humans. We're trying to build useful, and hopefully reliable, *tools* and *systems* for people.


P Reposted

The easiest way to "get ahead" in AI these days is to study a little bit of each data modality (text, vision, audio, tabular, time-series, etc.) If you're clever enough, you'll start to notice some tricks used in one modality that aren't used in others. A recent example I…


P Reposted

I don't know how I missed this but Style Aligned Image Generation by @Google Research is very much a meme material 🤩 You can try it here 👉 huggingface.co/spaces/fffilon… @fffiloni

Tweet Image 1

Loading...

Something went wrong.


Something went wrong.