@braintelligence.bsky.social
Believer in inclusive democracy
Posting mostly about AI/ML and tech if I can help it
Posting mostly about AI/ML and tech if I can help it
You can literally see the frontier of human knowledge
AELLA is an open-science initiative to make scientific research accessible via structured summaries created by LLMs
Available now:
- Dataset of 100K summaries
- 2 fine-tuned LLMs
- 3d visualizer
Available now:
- Dataset of 100K summaries
- 2 fine-tuned LLMs
- 3d visualizer
November 12, 2025 at 5:36 AM
You can literally see the frontier of human knowledge
Reposted
AELLA is an open-science initiative to make scientific research accessible via structured summaries created by LLMs
Available now:
- Dataset of 100K summaries
- 2 fine-tuned LLMs
- 3d visualizer
Available now:
- Dataset of 100K summaries
- 2 fine-tuned LLMs
- 3d visualizer
November 12, 2025 at 5:35 AM
AELLA is an open-science initiative to make scientific research accessible via structured summaries created by LLMs
Available now:
- Dataset of 100K summaries
- 2 fine-tuned LLMs
- 3d visualizer
Available now:
- Dataset of 100K summaries
- 2 fine-tuned LLMs
- 3d visualizer
The grok team should be ostracized from the ai community
It's now stating affirmatively the opposite of reality.
November 12, 2025 at 4:52 AM
The grok team should be ostracized from the ai community
Reposted
vLLM is now optimized for Arc Pro B-Series GPUs
You can buy 8 of these GPUs and cluster them for 192GB of VRAM for under $5,000.
blog.vllm.ai/2025/11/11/i...
You can buy 8 of these GPUs and cluster them for 192GB of VRAM for under $5,000.
blog.vllm.ai/2025/11/11/i...
Fast and Affordable LLMs serving on Intel Arc Pro B-Series GPUs with vLLM
Intel® Arc™ Pro B-Series GPU Family GPUs deliver powerful AI capabilities with a focus on accessibility and exceptional price-to-performance ratios. Their large memory capacity and scalability with mu...
blog.vllm.ai
November 11, 2025 at 11:21 PM
vLLM is now optimized for Arc Pro B-Series GPUs
You can buy 8 of these GPUs and cluster them for 192GB of VRAM for under $5,000.
blog.vllm.ai/2025/11/11/i...
You can buy 8 of these GPUs and cluster them for 192GB of VRAM for under $5,000.
blog.vllm.ai/2025/11/11/i...
Reposted
Actually naked-eye visible in Champaign
November 12, 2025 at 2:23 AM
Actually naked-eye visible in Champaign
Reposted
My political hot take is that US voters should be able to recall elected representatives via petition - at any time.
No more hiding behind election terms.
No more hiding behind election terms.
November 11, 2025 at 1:05 PM
My political hot take is that US voters should be able to recall elected representatives via petition - at any time.
No more hiding behind election terms.
No more hiding behind election terms.
apple brought back ipod socks but made them wildly overpriced:
www.apple.com/newsroom/202...
www.apple.com/newsroom/202...
Introducing iPhone Pocket: a beautiful way to wear and carry iPhone
ISSEY MIYAKE and Apple today unveiled iPhone Pocket, designed to fit any iPhone as well as all pocketable items.
www.apple.com
November 11, 2025 at 4:55 PM
apple brought back ipod socks but made them wildly overpriced:
www.apple.com/newsroom/202...
www.apple.com/newsroom/202...
Reposted
NEWS: The UK is no longer sharing intelligence with the US about suspected drug trafficking vessels in the Caribbean because it does not want to be complicit in US military strikes and believes the attacks are illegal, sources familiar with the matter told CNN. edition.cnn.com/2025/11/11/p...
Exclusive: UK suspends some intelligence sharing with US over boat strike concerns in major break | CNN Politics
The United Kingdom is no longer sharing intelligence with the US about suspected drug trafficking vessels in the Caribbean because it does not want to be complicit in US military strikes and believes ...
edition.cnn.com
November 11, 2025 at 3:06 PM
NEWS: The UK is no longer sharing intelligence with the US about suspected drug trafficking vessels in the Caribbean because it does not want to be complicit in US military strikes and believes the attacks are illegal, sources familiar with the matter told CNN. edition.cnn.com/2025/11/11/p...
Reposted
Good news: more than 50% of the world’s economies have seen carbon emissions from fossil fuel power generation peaking. And with the speed of change in the electricity sector many more will join this growing list in coming years.
November 11, 2025 at 7:18 AM
Good news: more than 50% of the world’s economies have seen carbon emissions from fossil fuel power generation peaking. And with the speed of change in the electricity sector many more will join this growing list in coming years.
Reposted
This paper received the perfect score for upcoming NeurIPS at San Diego or is it Mexico City?
November 10, 2025 at 10:38 PM
This paper received the perfect score for upcoming NeurIPS at San Diego or is it Mexico City?
Reposted
Baidu released ERNIE-4.5-VL-28B-A3B-Thinking, a lightweight multimodal reasoning model.
- 3B active parameters with enhanced semantic alignment between visual and language modalities
- "Thinking with Images" feature that enables zooming in and out to capture finer details
- Apache License 2.0
- 3B active parameters with enhanced semantic alignment between visual and language modalities
- "Thinking with Images" feature that enables zooming in and out to capture finer details
- Apache License 2.0
November 11, 2025 at 12:56 PM
Baidu released ERNIE-4.5-VL-28B-A3B-Thinking, a lightweight multimodal reasoning model.
- 3B active parameters with enhanced semantic alignment between visual and language modalities
- "Thinking with Images" feature that enables zooming in and out to capture finer details
- Apache License 2.0
- 3B active parameters with enhanced semantic alignment between visual and language modalities
- "Thinking with Images" feature that enables zooming in and out to capture finer details
- Apache License 2.0
Reposted
This seems like good news... www.theguardian.com/world/2025/n...
China’s CO2 emissions have been flat or falling for past 18 months, analysis finds
World’s biggest polluter on track to hit peak emissions target early but miss goal for cutting carbon intensity
www.theguardian.com
November 11, 2025 at 1:00 AM
This seems like good news... www.theguardian.com/world/2025/n...
Reposted
Space DJ turns genre embeddings into a playable galaxy—pilot a ship, the music follows. 🚀
Key stats
768→128 PCA compression; 3D UMAP projection; three.js rendering; autopilot drift; high‑dim neighbors surfacing hidden similarities.
Key stats
768→128 PCA compression; 3D UMAP projection; three.js rendering; autopilot drift; high‑dim neighbors surfacing hidden similarities.
November 11, 2025 at 3:03 PM
Space DJ turns genre embeddings into a playable galaxy—pilot a ship, the music follows. 🚀
Key stats
768→128 PCA compression; 3D UMAP projection; three.js rendering; autopilot drift; high‑dim neighbors surfacing hidden similarities.
Key stats
768→128 PCA compression; 3D UMAP projection; three.js rendering; autopilot drift; high‑dim neighbors surfacing hidden similarities.
Reposted
Meta just dropped a universal speech engine—Omnilingual ASR—built to hear the world.
Key stats
1600+ langs - 500 first-time served - 300M–7B params - w2v 2.0 (7B) backbone.
Key stats
1600+ langs - 500 first-time served - 300M–7B params - w2v 2.0 (7B) backbone.
November 11, 2025 at 2:35 PM
Meta just dropped a universal speech engine—Omnilingual ASR—built to hear the world.
Key stats
1600+ langs - 500 first-time served - 300M–7B params - w2v 2.0 (7B) backbone.
Key stats
1600+ langs - 500 first-time served - 300M–7B params - w2v 2.0 (7B) backbone.
Reposted
For those who keep asking me about 3I/ATLAS, here's a great writeup by my friend
@astrowright.bsky.social
sites.psu.edu/astrowright/...
@astrowright.bsky.social
sites.psu.edu/astrowright/...
Loeb’s 3I/ATLAS “Anomalies” Explained
Avi Loeb continues to claim that 3I/ATLAS has many anomalous behaviors that lead to the conclusion that it “might” be an alien spacecraft. He carefully hedges the probability that it is a spacecraft ...
sites.psu.edu
November 11, 2025 at 12:03 AM
For those who keep asking me about 3I/ATLAS, here's a great writeup by my friend
@astrowright.bsky.social
sites.psu.edu/astrowright/...
@astrowright.bsky.social
sites.psu.edu/astrowright/...
This is pretty funny. People just hanging out with a random person teleoperating their robot:
From the FunnyRobots community on Reddit: Rajesh really posted up like he owns the place
Explore this post and more from the FunnyRobots community
www.reddit.com
November 11, 2025 at 3:40 AM
This is pretty funny. People just hanging out with a random person teleoperating their robot:
Some pretty decent nanobanana2 examples of what it means that next gen LLMs will think/reason in other modalities:
From the singularity community on Reddit: Nano Banana 2 CRAZY image outputs
Explore this post and more from the singularity community
www.reddit.com
November 11, 2025 at 3:28 AM
Some pretty decent nanobanana2 examples of what it means that next gen LLMs will think/reason in other modalities:
Reposted
Along with Baguettotron we release the smallest viable language model to date. Monad, a 56M transformer, trained on the English part of SYNTH with non-random performance on MMLU. Desiging Monad an engineering challenge requiring a custom tiny tokenizer. huggingface.co/PleIAs/Monad
November 10, 2025 at 5:33 PM
Along with Baguettotron we release the smallest viable language model to date. Monad, a 56M transformer, trained on the English part of SYNTH with non-random performance on MMLU. Desiging Monad an engineering challenge requiring a custom tiny tokenizer. huggingface.co/PleIAs/Monad
Reposted
SYNTH is a collection of several synthetic playgrounds: data is not generated through simple prompts but by integrating smaller fine-tuned models into workflows with seeding, constraints, and formal verifications/checks.
November 10, 2025 at 5:31 PM
SYNTH is a collection of several synthetic playgrounds: data is not generated through simple prompts but by integrating smaller fine-tuned models into workflows with seeding, constraints, and formal verifications/checks.
Randomly turned on tv to a Star Trek TOS episode about an AI computer called M5 that goes rogue and takes over the Enterprise
November 11, 2025 at 1:33 AM
Randomly turned on tv to a Star Trek TOS episode about an AI computer called M5 that goes rogue and takes over the Enterprise
Reposted
Skills support for Letta Code is now available on release 0.2.0!
Use all of Anthropic's skills but with a stateful agent that never forgets.
Use all of Anthropic's skills but with a stateful agent that never forgets.
November 10, 2025 at 9:43 PM
Skills support for Letta Code is now available on release 0.2.0!
Use all of Anthropic's skills but with a stateful agent that never forgets.
Use all of Anthropic's skills but with a stateful agent that never forgets.
Reposted
Intel’s top AI executive leaves for OpenAI after 6 months in role-FT
November 11, 2025 at 1:11 AM
Intel’s top AI executive leaves for OpenAI after 6 months in role-FT
Reposted
The largest egocentric dataset.
Egocentric (first person) video is a general learning framework that passively captures how skilled workers do their jobs.
- 10,000 hours
- 2,153 factory workers
- 1,080,000,000 frames
Egocentric (first person) video is a general learning framework that passively captures how skilled workers do their jobs.
- 10,000 hours
- 2,153 factory workers
- 1,080,000,000 frames
November 10, 2025 at 11:54 PM
The largest egocentric dataset.
Egocentric (first person) video is a general learning framework that passively captures how skilled workers do their jobs.
- 10,000 hours
- 2,153 factory workers
- 1,080,000,000 frames
Egocentric (first person) video is a general learning framework that passively captures how skilled workers do their jobs.
- 10,000 hours
- 2,153 factory workers
- 1,080,000,000 frames
Reposted
Can LLMs accurately aggregate information over long context length, 128K for example.
No model achieves >50% accuracy at 128K!
They introduce Oolong, a dataset of simple-to-verify information aggregation questions over long inputs.
No model achieves >50% accuracy at 128K!
They introduce Oolong, a dataset of simple-to-verify information aggregation questions over long inputs.
November 10, 2025 at 10:59 PM
Can LLMs accurately aggregate information over long context length, 128K for example.
No model achieves >50% accuracy at 128K!
They introduce Oolong, a dataset of simple-to-verify information aggregation questions over long inputs.
No model achieves >50% accuracy at 128K!
They introduce Oolong, a dataset of simple-to-verify information aggregation questions over long inputs.
Reposted
Assumption: A token which never appears in the training data never receives updates during training and so does not move.
Data: lol no those fuckers be walkin
People, always check your assumptions. Even if you can't think of a way for dead tokens to move, that doesn't mean they don't move.
Data: lol no those fuckers be walkin
People, always check your assumptions. Even if you can't think of a way for dead tokens to move, that doesn't mean they don't move.
November 10, 2025 at 3:37 PM
Assumption: A token which never appears in the training data never receives updates during training and so does not move.
Data: lol no those fuckers be walkin
People, always check your assumptions. Even if you can't think of a way for dead tokens to move, that doesn't mean they don't move.
Data: lol no those fuckers be walkin
People, always check your assumptions. Even if you can't think of a way for dead tokens to move, that doesn't mean they don't move.