Telmo(@thetruetelmo) 's Twitter Profileg
Telmo

@thetruetelmo

ML Engineer and skeptic. Ex-New Yorker. Previously @Apple, @GoogleAI, @Unbabel

ID:1031605521982386176

calendar_today20-08-2018 18:14:57

1,4K Tweets

198 Followers

368 Following

Sabine Hossenfelder(@skdh) 's Twitter Profile Photo

How I fell out of love with academia

(this video was an accidental publication/scheduling blunder😬😬 but well uh, happy Friday I guess)

youtube.com/watch?v=LKiBlG…

account_circle
FFmpeg(@FFmpeg) 's Twitter Profile Photo

The xz fiasco has shown how a dependence on unpaid volunteers can cause major problems. Trillion dollar corporations expect free and urgent support from volunteers.

Microsoft MicrosoftTeams posted on a bug tracker full of volunteers that their issue is 'high priority'

The xz fiasco has shown how a dependence on unpaid volunteers can cause major problems. Trillion dollar corporations expect free and urgent support from volunteers. @Microsoft @MicrosoftTeams posted on a bug tracker full of volunteers that their issue is 'high priority'
account_circle
AK(@_akhaliq) 's Twitter Profile Photo

SaulLM-7B

A pioneering Large Language Model for Law

In this paper, we introduce SaulLM-7B, a large language model (LLM) tailored for the legal domain. With 7 billion parameters, SaulLM-7B is the first LLM designed explicitly for legal text comprehension and generation.

SaulLM-7B A pioneering Large Language Model for Law In this paper, we introduce SaulLM-7B, a large language model (LLM) tailored for the legal domain. With 7 billion parameters, SaulLM-7B is the first LLM designed explicitly for legal text comprehension and generation.
account_circle
Matthew Green(@matthew_d_green) 's Twitter Profile Photo

A thing I worry about in the (academic) privacy field is that our work isn’t really improving privacy globally. If anything it would be more accurate to say we’re finding ways to encourage the collection and synthesis of more data, by applying a thin veneer of local “privacy.”

account_circle
Michael Lynch(@deliberatecoder) 's Twitter Profile Photo

2023 was the best year yet for my indie hardware business. We reached $1M in revenue and $236k in profit. I learned how much low-latency responsibility was really costing me and how much the team benefits from async escalation. mtlynch.io/solo-developer…

account_circle
(((ل()(ل() 'yoav))))👾(@yoavgo) 's Twitter Profile Photo

It is very tempting to read into the 'grandmaster chess without search' result way beyond what's actually there. I was also mislead on first read, but further reading and reflection brought me down to earth. I wrote a bit about it.

gist.github.com/yoavg/8b98bbd7…

account_circle
Mahmoud Rasmi(@Decafquest) 's Twitter Profile Photo

The definitive guide to writing and presenting a top-notch philosophy paper.

The title:
Ignore what they might have told you about having to leave the title until the end. To get a grip on the paper’s spirit, you need to kick it off with a fancy-sounding, jaw-dropping title that…

The definitive guide to writing and presenting a top-notch philosophy paper. The title: Ignore what they might have told you about having to leave the title until the end. To get a grip on the paper’s spirit, you need to kick it off with a fancy-sounding, jaw-dropping title that…
account_circle
Haitham Bou Ammar (هيثم بو عمار)(@hbouammar) 's Twitter Profile Photo

With this hype around LLMs, I am seeing tons of well respected professors publishing half (if not quarter) - baked work just to keep up with the competition!

I will advice them as they did advise me once during my PhD. This has to stop, focus on quality and not quantity… I feel…

account_circle
Mahmoud Rasmi(@Decafquest) 's Twitter Profile Photo

The bureaucracy law:

For each application you present, no matter how complete it is, there will always remain one document missing.

account_circle
Mahmoud Rasmi(@Decafquest) 's Twitter Profile Photo

📢 The book is now live! Grab your copy today.

Philosophy for Business Leaders: Asking Questions, Navigating Uncertainty, and the Quest for Meaning.

Available in paperback and Kindle editions.

amazon.com/dp/8409559501

📢 The book is now live! Grab your copy today. Philosophy for Business Leaders: Asking Questions, Navigating Uncertainty, and the Quest for Meaning. Available in paperback and Kindle editions. amazon.com/dp/8409559501
account_circle
Official Jon Lajoie(@jonlajoiecomedy) 's Twitter Profile Photo

Happy Halloween! This is a fantastic fan-made video for my Wolfie's Just Fine song 'Get Away' made by David Kirkbride David Kirkbride. It's an edit of a bunch of 80s horror movies (Can you name them all?). It captures the essence of the song and is perfect for Halloween, enjoy!

account_circle
Mahmoud Rasmi(@Decafquest) 's Twitter Profile Photo

Coming out in December.

Philosophy for Business Leaders, Asking Questions, Navigating Uncertainty, and the Quest for Meaning

Coming out in December. Philosophy for Business Leaders, Asking Questions, Navigating Uncertainty, and the Quest for Meaning
account_circle
Cory Doctorow NONCONSENSUAL BLUE TICK(@doctorow) 's Twitter Profile Photo

has no cannie adversary than Apple, whose most innovative work is dreaming up new ways to sabotage electronics repair while claiming to be a caring environmental steward, a lie to cover the mountains of that Apple dooms our descendants to wade through.

1/

#RightToRepair has no cannie adversary than @Apple, whose most innovative work is dreaming up new ways to sabotage electronics repair while claiming to be a caring environmental steward, a lie to cover the mountains of #ewaste that Apple dooms our descendants to wade through. 1/
account_circle
Birchlabs(@Birchlabs) 's Twitter Profile Photo

One Wide Feedforward is All You Need
Apple is *not* sleeping on LLM training.
in encoder-decoder models, you can
- remove the decoder's FFN
- tie the weights in encoder FFNs (but make it wider)
improves accuracy and latency *and* reduces params!
arxiv.org/abs/2309.01826

One Wide Feedforward is All You Need Apple is *not* sleeping on LLM training. in encoder-decoder models, you can - remove the decoder's FFN - tie the weights in encoder FFNs (but make it wider) improves accuracy and latency *and* reduces params! arxiv.org/abs/2309.01826
account_circle