Our mission is to learn from evolution and redefine computing, using living neurons to make silicon-based AI models more stable, scalable and efficient.
San Francisco, California
- Tweets 64
- Following 103
- Followers 705
- Likes 39
Pinned Tweet
How do we make AI models better without making them bigger?
At TBC, we study how neurons solve problems to make AI video models better, faster and cheaper.
See why we started and what we're building: bit.ly/3Sf4qXM
3
11
2,452
How do we make AI models better without making them bigger?
At TBC, we study how neurons solve problems to make AI video models better, faster and cheaper.
See why we started and what we're building: bit.ly/3Sf4qXM
3
11
2,452
I was going through all the information about the The Bio Computing from different resource like Rat neuron, also through the Ted talks of the Mark Solms, David Chalmers and also I was researching about this from few months, would love to know about any research opportunities.
1
1
19
Check out our blog for more of the research: tbc.co/blog
2
1
45
The next leap in AI is not another warehouse of GPUs.
It will be Biological Computing.
TBC grows hundreds of thousands of neurons on a 4,096-electrode dish.
They encode an image or a token as an electrical pattern, send it into the culture and record how the cells process it across space and time.
That is not a metaphor for a neural net. It is the wet hardware.
The point is efficiency.
A fruit fly brain does useful compute on about 140,000 neurons and less than 6 mWh a day. Silicon does not.
That is why DARPA stood up O-CIRCUIT this year. The program is building biological processing units for edge learning and inference, where power is scarce and a diesel generator is not an option.
Same problem TBC is working in the lab: stop throwing more electricity at the model, start using the compute biology already solved.
@BioComputingCo first product is software derived from those recordings, not a brain in the rack.
"Whoever Has The Best AI, Will Rule The World" From Episode 72 with @shadmoney_
2
2
16
1,774
Real brain cells. 4,096 electrodes. Neurons firing.
Come inside TBC with @awscloud piped.video/watch?v=R4eKRihw…
1
4
10
629
What happens when neurons meet silicon? A new world emerges.
Partnering with Amazon to bring it to customers. DM me for early access.
18
16
138
92,910
Did you know you can learn interesting things from computations in real rat neurons? I love what @AlexKsendzovsky and team are doing at @BioComputingCo are doing.
What happens when neurons meet silicon? A new world emerges.
Partnering with Amazon to bring it to customers. DM me for early access.
23
64
643
93,285
Living brain cells use a minimum amount of energy for computations. If we combine them with chips, AI will become significantly cheaper to run.
2
3
296
Did you know you can learn interesting things from computations in real rat neurons? I love what @AlexKsendzovsky and team are doing at @BioComputingCo are doing.
What happens when neurons meet silicon? A new world emerges.
Partnering with Amazon to bring it to customers. DM me for early access.
23
64
643
93,285
From the dish to cloud.
🧠 @BioComputingCo records how living neurons process images and turns those patterns into AI software. The result is video generation that’s up to 5x faster and significantly cheaper to run, now available to select AWS customers.
@LaurenGoode reports for @WIRED: go.aws/4xESNrW
1
1
8
1,180
The Biological Computing Company is bringing its AI tools to Amazon Web Services in a major boost for a once-fringe field that aims to marry nature with code. wired.com/story/ai-models-bu…
6
11
25
23,620
The Biological Computing Company is bringing its AI tools to Amazon Web Services in a major boost for a once-fringe field that aims to marry nature with code. wired.com/story/ai-models-bu…
6
11
25
23,620
Thanks for stopping by our lab @LaurenGoode and sharing what we're up to @WIRED
2
83
this is just the beginning…
Nerd alert: AI models built on rat brain neurons—which are placed on multi-electrode silicon arrays, respond to electrical signals, get translated into code, and are then used to train models—are that much closer to reality. As in, they're now available through AWS wired.com/story/ai-models-bu…
6
2
18
4,082
@LLMJunky this is what I'm talking about!
1
1
175
AI has trained us to expect a tradeoff:
Faster = worse
Better = more expensive
What if biology gives us another option?
We’re seeing better output with less compute.
@AlexKsendzovsky gets into why here ↓
piped.video/watch?v=BjW-yEQy…
7
909
Gotta love @Figure_robot
At its San Jose headquarters it has a bunch of robots walking around with no nerds watching them.
199
140
1,544
891,541
@Scobleizer come to our lab in SF. You'll see how we use real neurons to make AI more efficient. It's pretty mind blowing!
1
21
Today is my last day as Editorial Director of
@VentureBeat
My first story started with 5,000 people (and a few llamas) at Hugging Face’s “Woodstock of AI.” My last ends with Nvidia buying them for $12.9B.
700+ stories later, the circle is complete.
Reflections on three years at the frontier — and a hint at the next vantage point:
venturebeat.com/technology/m…
26
4
134
8,383
Congrats on a great run @MichaelFNunez!
1
45
What can real neurons teach AI?
At TBC, we turn discoveries from real neurons into better AI video: faster generation, better performance, better representation of the physical world.
@AlexKsendzovsky explains the connection in under 60 seconds ↓
1
5
487
Full piece here: eejournal.com/fish_fry/bio-i…
1
310
This is probably the fastest way to understand TBC.
@atlasberry008 managed to explain @BioComputingCo in 2 minutes. i’ve been trying for years.
instagram.com/reel/DdCY4idB1…
1
2
421
We use real neurons to improve AI.
Now we’re launching the world’s first neuron-derived video generation model.
Faster. Cheaper. Better.
Running on silicon. Upgraded by biology.
Get early access: tbc.co/early-access
1
2
8
1,532
