Dr S Osindero
@sindero.bsky.social
Jack of all trades, master of some
•ⒶⒾ@DeepMind
•Personal account
•Mostly AI/ML/CS/biotech/physics/math/complexity/art/animals
•They/Any
•⚛️🏳️🌈🇬🇧🇨🇦🇺🇸🇪🇺
•ⒶⒾ@DeepMind
•Personal account
•Mostly AI/ML/CS/biotech/physics/math/complexity/art/animals
•They/Any
•⚛️🏳️🌈🇬🇧🇨🇦🇺🇸🇪🇺
Reposted by Dr S Osindero
This is the most interesting contribution to the discussion about consciousness I've read in years. For me, the discussion is pretty much resolved.
Why does the concept of 'red' not feel like the experience of red? Our paper argues this disconnect exists because our deep-seated need to be understood will always exceed the limits of the language we invent to achieve it.
arxiv.org/pdf/2506.12086
#Consciousness #Neuroscience #Philosophy #AI
arxiv.org/pdf/2506.12086
#Consciousness #Neuroscience #Philosophy #AI
arxiv.org
June 17, 2025 at 3:32 PM
This is the most interesting contribution to the discussion about consciousness I've read in years. For me, the discussion is pretty much resolved.
Reposted by Dr S Osindero
Gemma 3 explained: Longer context, image support, and a new 1B model. → goo.gle/4lV8iaw
Other key enhancements:
🔸 Best model that fits in a single consumer GPU or TPU host
🔸 KV-cache memory reduction with 5-to-1 interleaved attention
🔸 And more!
Read the blog for the full details on Gemma 3.
Other key enhancements:
🔸 Best model that fits in a single consumer GPU or TPU host
🔸 KV-cache memory reduction with 5-to-1 interleaved attention
🔸 And more!
Read the blog for the full details on Gemma 3.
Gemma explained: What’s new in Gemma 3- Google Developers Blog
Google's Gemma 3 model includes vision-language support and architectural changes for resource-friendly multimodal language models.
goo.gle
April 30, 2025 at 9:46 PM
Gemma 3 explained: Longer context, image support, and a new 1B model. → goo.gle/4lV8iaw
Other key enhancements:
🔸 Best model that fits in a single consumer GPU or TPU host
🔸 KV-cache memory reduction with 5-to-1 interleaved attention
🔸 And more!
Read the blog for the full details on Gemma 3.
Other key enhancements:
🔸 Best model that fits in a single consumer GPU or TPU host
🔸 KV-cache memory reduction with 5-to-1 interleaved attention
🔸 And more!
Read the blog for the full details on Gemma 3.
Reposted by Dr S Osindero
Why does the concept of 'red' not feel like the experience of red? Our paper argues this disconnect exists because our deep-seated need to be understood will always exceed the limits of the language we invent to achieve it.
arxiv.org/pdf/2506.12086
#Consciousness #Neuroscience #Philosophy #AI
arxiv.org/pdf/2506.12086
#Consciousness #Neuroscience #Philosophy #AI
arxiv.org
June 17, 2025 at 1:52 PM
Why does the concept of 'red' not feel like the experience of red? Our paper argues this disconnect exists because our deep-seated need to be understood will always exceed the limits of the language we invent to achieve it.
arxiv.org/pdf/2506.12086
#Consciousness #Neuroscience #Philosophy #AI
arxiv.org/pdf/2506.12086
#Consciousness #Neuroscience #Philosophy #AI
Reposted by Dr S Osindero
💯
Hallucination is totally the wrong word, implying it is perceiving the world incorrectly.
But it's generating false, plausible sounding statements. Confabulation is literally the perfect word.
So, let's all please start referring to any junk that an LLM makes up as "confabulations".
Hallucination is totally the wrong word, implying it is perceiving the world incorrectly.
But it's generating false, plausible sounding statements. Confabulation is literally the perfect word.
So, let's all please start referring to any junk that an LLM makes up as "confabulations".
petition to change the word describing ChatGPT's mistakes from 'hallucinations' to 'confabulations'
A hallucination is a false subjective sensory experience. ChatGPT doesn't have experiences!
It's just making up plausible-sounding bs, covering knowledge gaps. That's confabulation
A hallucination is a false subjective sensory experience. ChatGPT doesn't have experiences!
It's just making up plausible-sounding bs, covering knowledge gaps. That's confabulation
December 11, 2024 at 2:47 PM
💯
Hallucination is totally the wrong word, implying it is perceiving the world incorrectly.
But it's generating false, plausible sounding statements. Confabulation is literally the perfect word.
So, let's all please start referring to any junk that an LLM makes up as "confabulations".
Hallucination is totally the wrong word, implying it is perceiving the world incorrectly.
But it's generating false, plausible sounding statements. Confabulation is literally the perfect word.
So, let's all please start referring to any junk that an LLM makes up as "confabulations".
Reposted by Dr S Osindero
Very happy to announce the publication of our latest paper:
A theory of appropriateness with applications to generative artificial intelligence
arxiv.org/abs/2412.19010
And happy new year everyone!
A theory of appropriateness with applications to generative artificial intelligence
arxiv.org/abs/2412.19010
And happy new year everyone!
A theory of appropriateness with applications to generative artificial intelligence
What is appropriateness? Humans navigate a multi-scale mosaic of interlocking notions of what is appropriate for different situations. We act one way with our friends, another with our family, and yet...
arxiv.org
December 31, 2024 at 7:48 AM
Very happy to announce the publication of our latest paper:
A theory of appropriateness with applications to generative artificial intelligence
arxiv.org/abs/2412.19010
And happy new year everyone!
A theory of appropriateness with applications to generative artificial intelligence
arxiv.org/abs/2412.19010
And happy new year everyone!
Reposted by Dr S Osindero
I can't believe they've just cancelled the Epidemic Intelligence Service program at CDC.
This program trains the best & brightest epidemiologists, who then go on to have distinguished careers in public health, serving at CDC, in state health departments, overseas, ...
This program trains the best & brightest epidemiologists, who then go on to have distinguished careers in public health, serving at CDC, in state health departments, overseas, ...
February 15, 2025 at 5:56 AM
I can't believe they've just cancelled the Epidemic Intelligence Service program at CDC.
This program trains the best & brightest epidemiologists, who then go on to have distinguished careers in public health, serving at CDC, in state health departments, overseas, ...
This program trains the best & brightest epidemiologists, who then go on to have distinguished careers in public health, serving at CDC, in state health departments, overseas, ...
Reposted by Dr S Osindero
The idea of "AI alignment" grew out of a community that thought you could solve morals like it was a CS problem set. Nice to see a more nuanced take.
Announcing our latest arxiv paper:
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
arxiv.org/abs/2505.05197
We argue for a view of AI safety centered on preventing disagreement from spiraling into conflict.
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
arxiv.org/abs/2505.05197
We argue for a view of AI safety centered on preventing disagreement from spiraling into conflict.
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
Artificial Intelligence (AI) systems are increasingly placed in positions where their decisions have real consequences, e.g., moderating online spaces, conducting research, and advising on policy. Ens...
arxiv.org
May 9, 2025 at 2:49 PM
The idea of "AI alignment" grew out of a community that thought you could solve morals like it was a CS problem set. Nice to see a more nuanced take.
Reposted by Dr S Osindero
Concordia is a library for generative agent-based modeling that works like a table-top role-playing game.
It's open source and model agnostic.
Try it today!
github.com/google-deepm...
It's open source and model agnostic.
Try it today!
github.com/google-deepm...
GitHub - google-deepmind/concordia: A library for generative social simulation
A library for generative social simulation. Contribute to google-deepmind/concordia development by creating an account on GitHub.
github.com
November 16, 2024 at 11:49 PM
Concordia is a library for generative agent-based modeling that works like a table-top role-playing game.
It's open source and model agnostic.
Try it today!
github.com/google-deepm...
It's open source and model agnostic.
Try it today!
github.com/google-deepm...
Reposted by Dr S Osindero
Announcing our latest arxiv paper:
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
arxiv.org/abs/2505.05197
We argue for a view of AI safety centered on preventing disagreement from spiraling into conflict.
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
arxiv.org/abs/2505.05197
We argue for a view of AI safety centered on preventing disagreement from spiraling into conflict.
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
Artificial Intelligence (AI) systems are increasingly placed in positions where their decisions have real consequences, e.g., moderating online spaces, conducting research, and advising on policy. Ens...
arxiv.org
May 9, 2025 at 11:39 AM
Announcing our latest arxiv paper:
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
arxiv.org/abs/2505.05197
We argue for a view of AI safety centered on preventing disagreement from spiraling into conflict.
Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt
arxiv.org/abs/2505.05197
We argue for a view of AI safety centered on preventing disagreement from spiraling into conflict.
Reposted by Dr S Osindero
Authors:
Viorica Pătrăucean, Xu Owen He, Joseph Heyward, Chuhan Zhang, Mehdi S. M. Sajjadi, George-Cristian Muraru, Artem Zholus, Mahdi Karami, Ross Goroshin, Yutian Chen, Simon Osindero, João Carreira, Razvan Pascanu
Original post:
www.linkedin.com/posts/vioric...
Viorica Pătrăucean, Xu Owen He, Joseph Heyward, Chuhan Zhang, Mehdi S. M. Sajjadi, George-Cristian Muraru, Artem Zholus, Mahdi Karami, Ross Goroshin, Yutian Chen, Simon Osindero, João Carreira, Razvan Pascanu
Original post:
www.linkedin.com/posts/vioric...
Viorica Patraucean on LinkedIn: Super excited to share our recent work on designing more efficient video…
Super excited to share our recent work on designing more efficient video models: TRecViT https://lnkd.in/ehh4gGbn alternates SSM blocks (LRUs) that integrate…
www.linkedin.com
January 10, 2025 at 3:44 PM
Authors:
Viorica Pătrăucean, Xu Owen He, Joseph Heyward, Chuhan Zhang, Mehdi S. M. Sajjadi, George-Cristian Muraru, Artem Zholus, Mahdi Karami, Ross Goroshin, Yutian Chen, Simon Osindero, João Carreira, Razvan Pascanu
Original post:
www.linkedin.com/posts/vioric...
Viorica Pătrăucean, Xu Owen He, Joseph Heyward, Chuhan Zhang, Mehdi S. M. Sajjadi, George-Cristian Muraru, Artem Zholus, Mahdi Karami, Ross Goroshin, Yutian Chen, Simon Osindero, João Carreira, Razvan Pascanu
Original post:
www.linkedin.com/posts/vioric...
Reposted by Dr S Osindero
TRecViT: A Recurrent Video Transformer
arxiv.org/abs/2412.14294
Causal, 3× fewer parameters, 12× less memory, 5× higher FLOPs than (non-causal) ViViT, matching / outperforming on Kinetics & SSv2 action recognition.
Code and checkpoints out soon.
arxiv.org/abs/2412.14294
Causal, 3× fewer parameters, 12× less memory, 5× higher FLOPs than (non-causal) ViViT, matching / outperforming on Kinetics & SSv2 action recognition.
Code and checkpoints out soon.
January 10, 2025 at 3:44 PM
TRecViT: A Recurrent Video Transformer
arxiv.org/abs/2412.14294
Causal, 3× fewer parameters, 12× less memory, 5× higher FLOPs than (non-causal) ViViT, matching / outperforming on Kinetics & SSv2 action recognition.
Code and checkpoints out soon.
arxiv.org/abs/2412.14294
Causal, 3× fewer parameters, 12× less memory, 5× higher FLOPs than (non-causal) ViViT, matching / outperforming on Kinetics & SSv2 action recognition.
Code and checkpoints out soon.
Reposted by Dr S Osindero
WTF, Chris Onstad was making an Achewood cartoon for Netflix with Pendleton Ward, and the bastards canceled it?
https://www.theverge.com/23713940/achewood-chris-onstad-return-patreon-ai-raybot
https://www.theverge.com/23713940/achewood-chris-onstad-return-patreon-ai-raybot
Achewood is back, and it’s weirder than ever
A return from Circumstances.
www.theverge.com
May 10, 2023 at 2:14 AM
WTF, Chris Onstad was making an Achewood cartoon for Netflix with Pendleton Ward, and the bastards canceled it?
https://www.theverge.com/23713940/achewood-chris-onstad-return-patreon-ai-raybot
https://www.theverge.com/23713940/achewood-chris-onstad-return-patreon-ai-raybot
Reposted by Dr S Osindero
Actually that's a pretty good argument for a reboot.
May 10, 2023 at 3:47 AM
Actually that's a pretty good argument for a reboot.
Reposted by Dr S Osindero
Twitter is SOOO thirsty.
#DeadTwitter
#DeadTwitter
May 10, 2023 at 4:35 AM
Twitter is SOOO thirsty.
#DeadTwitter
#DeadTwitter
Reposted by Dr S Osindero
1-bit Hokusai’s ”The Great Wave”
May 10, 2023 at 5:07 AM
1-bit Hokusai’s ”The Great Wave”
Reposted by Dr S Osindero
I am maintaining three different narratives across Twitter, Mastodon, and Bluesky this is not sustainable. 😵💫
May 10, 2023 at 7:21 AM
I am maintaining three different narratives across Twitter, Mastodon, and Bluesky this is not sustainable. 😵💫
Reposted by Dr S Osindero
useful links for anyone wanting to build on the protocol & play with some of the primitives:
https://github.com/bluesky-social/atproto/tree/main/packages
https://www.npmjs.com/package/@atproto/api?activeTab=readme
https://www.npmjs.com/search?q=%40atproto
https://github.com/bluesky-social/atproto/tree/main/packages
https://www.npmjs.com/package/@atproto/api?activeTab=readme
https://www.npmjs.com/search?q=%40atproto
April 11, 2023 at 8:41 PM
useful links for anyone wanting to build on the protocol & play with some of the primitives:
https://github.com/bluesky-social/atproto/tree/main/packages
https://www.npmjs.com/package/@atproto/api?activeTab=readme
https://www.npmjs.com/search?q=%40atproto
https://github.com/bluesky-social/atproto/tree/main/packages
https://www.npmjs.com/package/@atproto/api?activeTab=readme
https://www.npmjs.com/search?q=%40atproto
I have ruined
the Twitter
that was in
regular use globally.
And which
you were probably
saving
for Freedom of Speech.
Forgive me
it was so tempting
to my fragile ego.
-Space K, 2022
the Twitter
that was in
regular use globally.
And which
you were probably
saving
for Freedom of Speech.
Forgive me
it was so tempting
to my fragile ego.
-Space K, 2022
April 11, 2023 at 9:21 PM
I have ruined
the Twitter
that was in
regular use globally.
And which
you were probably
saving
for Freedom of Speech.
Forgive me
it was so tempting
to my fragile ego.
-Space K, 2022
the Twitter
that was in
regular use globally.
And which
you were probably
saving
for Freedom of Speech.
Forgive me
it was so tempting
to my fragile ego.
-Space K, 2022