Picture

David Bartram-Shaw (DBS) @dbs_dsml

Chief AI Officer at @wearemesh_ai | Posting #AIinsights #AIopinions #AItools #RealWorldAI. Opinions are my own, no BS, I credit ideas that are not my own. London, UK Joined November 2021
  • Tweets
    150
  • Followers
    103
  • Following
    336
  • Likes
    665
David Bartram-Shaw (DBS) retweeted
karpathy Profile Picture

Andrej Karpathy @karpathy

a year ago

Awesome and highly useful: FineWeb-Edu 📚👏 High quality LLM dataset filtering the original 15 trillion FineWeb tokens to 1.3 trillion of the highest (educational) quality, as judged by a Llama 3 70B. +A highly detailed paper. Turns out that LLMs learn a lot better and faster…

gui_penedo Profile Picture

Guilherme Penedo @gui_penedo

a year ago

Awesome and highly useful: FineWeb-Edu 📚👏 High quality LLM dataset filtering the original 15 trillion FineWeb tokens to 1.3 trillion of the highest (educational) quality, as judged by a Llama 3 70B. +A highly detailed paper. Turns out that LLMs learn a lot better and faster… https://t.co/9nXaet5tmG

David Bartram-Shaw (DBS) retweeted
francoisfleuret Profile Picture

François Fleuret @francoisfleuret

a year ago

TL;DR: We designed a diffusion world model. Frame prediction is more accurate and improves sota on Atari for an RL agent trained with 100k frames. @EloiAlonso1, @AdamJelley2, @micheli_vincent, Anssi Kanervisto, @AmosStorkey, @Tea_Pearce @sciences_UNIGE @MSFTResearch 1/2

_akhaliq Profile Picture

AK @_akhaliq

a year ago

TL;DR: We designed a diffusion world model. Frame prediction is more accurate and improves sota on Atari for an RL agent trained with 100k frames. @EloiAlonso1, @AdamJelley2, @micheli_vincent, Anssi Kanervisto, @AmosStorkey, @Tea_Pearce @sciences_UNIGE @MSFTResearch 1/2

David Bartram-Shaw (DBS) retweeted
boazbaraktcs Profile Picture

Boaz Barak @boazbaraktcs

2 years ago

My view is that what makes super-alignment "super" is ensuring we can safely scale the capabilities of AIs even though we can't scale their human supervisors. For this, it is imperative to study the "weak teacher strong student" setting. Paper shows great promise in this area!

_akhaliq Profile Picture

AK @_akhaliq

2 years ago

My view is that what makes super-alignment "super" is ensuring we can safely scale the capabilities of AIs even though we can't scale their human supervisors. For this, it is imperative to study the "weak teacher strong student" setting. Paper shows great promise in this area!

David Bartram-Shaw (DBS) retweeted
iamtrask Profile Picture

⿻ Andrew Trask @iamtrask

2 years ago

LLMs believe every datapoint they see with 100% conviction. A LLM never says, "this doesn't make sense... let me exclude it from my training data". Everything is taken as truth. It is actually worse than this. Because of how perplexity/SGD/backprop works, datapoints which…

dbs_dsml Profile Picture

David Bartram-Shaw (DBS) @dbs_dsml

2 years ago

#AIopinions Such a honest and true viewpoint on what’s happening in AI right now from @natolambert 👌 A great discussion on the relationship between research and industry waves. Every person working in the field will empathize with this in depth view and take learnings from it.

natolambert Profile Picture

Nathan Lambert @natolambert

2 years ago

#AIopinions Such a honest and true viewpoint on what’s happening in AI right now from @natolambert 👌 A great discussion on the relationship between research and industry waves. Every person working in the field will empathize with this in depth view and take learnings from it.

David Bartram-Shaw (DBS) retweeted
ylecun Profile Picture

Yann LeCun @ylecun

3 years ago

Data on the intellectual contribution to AI from various research organizations. Some of organizations publish knowledge and open-source code for the entire world to use. Others just consume it.

johnjnay Profile Picture

John Nay @johnjnay

3 years ago

Data on the intellectual contribution to AI from various research organizations. Some of organizations publish knowledge and open-source code for the entire world to use. Others just consume it.

dbs_dsml Profile Picture

David Bartram-Shaw (DBS) @dbs_dsml

3 years ago

#aitools BOOM 💥 Here we go ...! 🚀 "With Azure OpenAI Service now generally available, more businesses can apply for access to the most advanced AI models in the world—including GPT-3.5, Codex, and DALL•E 2" azure.microsoft.com/en-us/blog/gen… #AzureOpenAI #Microsoft #ai #ml #DataScience

Trends for United States

Jake Moody
10,9 B posts
Lions
88,4 B posts
Packers
58,6 B posts
Bo Nix
6.089 posts
Jets
107 B posts
Steelers
118 B posts
49ers
46,8 B posts
Browns
79,4 B posts
Callahan
4.330 posts
Falcons
54,8 B posts
Dolphins
70,5 B posts
#GoPackGo
9.620 posts
Kittle
7.250 posts
Niners
5.172 posts
#VMAs
163 B posts
Jordan Love
6.837 posts
Broncos
36,9 B posts
Giants
104 B posts
Goff
5.888 posts
Titans
41,6 B posts

You might like

📊 Learning Data Science https://t.co/UWzMrvIsEZ
Rajguru Bhosale
@BhosaleRajguru
274 Followers
ai researcher @disneyresearch - prev applied scientist at @amazonscience (aws ai labs & alexa ai). tweets = personal opinions.
Guglielmo Camporese
@gucamporese
139 Followers
CEO and founder of @arcee_ai | @huggingface 🤗 alum. AI and Data Obsessed. Fitness Fanatic. Tattoo Enthusiast.
Mark McQuade
@mmcquade_ai_u
533 Followers
Ph.D. student at the Chinese University of Hong Kong, Shenzhen. Amphion co-founder. Interested in speech, music, and multi-modality.
Xueyao Zhang
@xueyao_98
120 Followers
An AI Researcher, 3x Kaggler | Creative Technologist | AI Product
Elvin Aghammadzada
@aghammadzada
90 Followers
Sharing Guides to Build Ridiculously Good AI Apps 💪🧠 | Generative AI | Python | 🔟+ yrs of AI expertise🛠🤖 | Connect if you’re technical🧑‍💻
AI Architect 🤖🔧
@ShivyanRiaz
970 Followers
Finding signal in the noise - generalist engineer
a|ex
@alexnhaus
165 Followers
🌟 ON THE JOB MARKET | AI Ph.D @UCRiverside
Hengyue Liu
@hengyueliu
33 Followers