Sébastien Toth (@tothsebastien) 's Twitter Profile
Sébastien Toth

@tothsebastien

ID: 1514990361613062147

calendar_today15-04-2022 15:34:03

906 Tweet

937 Takipçi

932 Takip Edilen

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

What's great about Twitter is that it shows how the communication model has become a dialogue rather than a monologue We're still falling into ad-like posts of course, yet the trend is clear

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Creating a start-up is about iterating long enough to succeed without going completely insane in the process 🤪 If you don't manage to do it, please at least #buildinpublic so you can join other misfits like us🤪

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Yes #StableDiffusion2 isn't revolutionary compared to v1. cudos to #stability because they share models to make all of us smarter instead of trying to hit the big news only

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

I've met with several successful entrepreneurs who've created companies worth 100s of millions of $. They were all crazy & 100% obsessed. Entrepreneurship is a drug 🤪 and I'm both admiring & scared by the amount of obsession these guys have It's a thin line !

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

YouTube made the concept of universities totally obsolete. Why pay hundreds of thousands of dollars for a mostly average course when you can freely access the best professors on any topic directly from your computer ? It's a matter of years before total obsolescence

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Theoretically, #twitter customers are corps buying ads. That means #users are actually the product. From a product management point of view, looks like Elon Musk is trying to put back #users as the main customers. Without discussing the "how", I find the "why" great.

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

It would be great for the scientific community to have the total compute budget spent (in term of AWS-equivalent pricing GPU hours) for each papers & model.

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

The worst thing about the 9-5 job is the tyranny of low expectations. For many it's becoming unbearable. Having hard challenges & being responsible to overcome them is part of what makes entrepreneurship so rewarding Let me fail, as long as I fought properly

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Interestingly, #ChatGPT is behind fine-tuned #SOTA in most tasks. Of course, those #benchmarks are not perfectly designed, and metrics are flawed. But what's interesting is that ChatGPT has been trained to... look like it's helping us. And we love that.

Interestingly, #ChatGPT is behind fine-tuned #SOTA in most tasks.

Of course, those #benchmarks are not perfectly designed, and metrics are flawed.

But what's interesting is that ChatGPT has been trained to... look like it's helping us. 

And we love that.
Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

prediction: LLM-generated content will rapidly become worthless and low-value compaired to human-generated content (except for synthesis). We use language to dialog with other humans, it's losing value if there is no empathy towards the writer or the reader. Happened with chess

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

I think people don't get how much ChatGPT & its successors will significantly raise the worldwide average human intelligence over time. StockFish & its predecessor was a huge improvement for chess performance, same here.

Hervé "pyannote" Bredin (@hbredin) 's Twitter Profile Photo

#pyannote models are now #2 and #3 most downloaded audio models on Hugging Face hub. Yet, I am struggling finding applicants for the two open positions to make it even better (2 years research engineer, 3 years PhD scholarshop) Can anyone help? x.com/hbredin/status…

#pyannote models are now #2 and #3 most downloaded audio models on <a href="/huggingface/">Hugging Face</a>  hub.

Yet, I am struggling finding applicants for the two open positions to make it even better (2 years research engineer, 3 years PhD scholarshop)

Can anyone help? 

x.com/hbredin/status…
Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

I may be too old school but I think these kind of content have no place in an "academic research paper" Looks like that "papers" have become yet another form of content no different than twitter threads or medium posts. h-index is losing value, fast.

I may be too old school but I think these kind of content have no place in an "academic research paper"

Looks like that "papers" have become yet another form of content no different than twitter threads or medium posts.

h-index is losing value, fast.
Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Remember when you needed 6 months of research & data labeling to learn one simple task such as text classification or some kind of context-based generation ? That was 6 months ago. Now it takes a few minutes of prompt-tuning and you're good to go.

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Google's "We don't have any MOTA" leak made quite a lot of noise. People keep complaining about OpenAI, Anthropic & Google switch to closer & closer source. It's not about them getting open, it's about giving strength to open-source players #huggingface

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

Med-PaLM-2 outsandingly beating GPT3.5 shows a very interesting yet under-developed domain for research & business opportunities: Expert LLMs

Med-PaLM-2 outsandingly beating GPT3.5 shows a very interesting yet under-developed domain for research &amp; business opportunities:

Expert LLMs
Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

2023 is an amazing time for entrepreneurship with so many new use-cases unlocked by A.I. Funny that it's also this specific year many VC decided to buckle up and get tight on investments I feel like BAs are taking back a huge role on seeding start ups and that's pretty good

Sébastien Toth (@tothsebastien) 's Twitter Profile Photo

So #GPT-3.5-turbo has 20B params ? Didn't know that, extracted from this paper: arxiv.org/pdf/2310.17680… from Microsoft Research.

So #GPT-3.5-turbo has 20B params ? 

Didn't know that, extracted from this paper: arxiv.org/pdf/2310.17680… from Microsoft Research.