TBC Unveils AI Demo With 2x Video Quality and 4.4x Lower Inference Cost
Updated
Updated · PR Newswire · Jul 30
TBC Unveils AI Demo With 2x Video Quality and 4.4x Lower Inference Cost
2 articles · Updated · PR Newswire · Jul 30
Summary
The Biological Computing Co. released a public proof of concept that applies its neurally optimized software to Decart's OASIS 500M interactive video model, offering a playable demo on its website.
Matched external testing with Bluesky Compute found the optimized model doubled a selected video-quality benchmark, cut inference cost by about 4.4 times and generated more than three times as much coherent video as the base model.
TBC said the demo is its first working application of optimization methods derived from experiments on living neural networks, then translated into lightweight software that runs on standard GPU and cloud infrastructure.
Later this year, the San Francisco company plans to use the same approach in a larger text-to-video model, which it intends to launch as its first commercial product and is already opening for beta interest.
With tech giants racing for neuro-AI dominance, will TBC’s brain-inspired software actually scale enough to dethrone heavyweight video models this year?
Can living brain cells truly unlock the secret to slashing AI costs, or is this biological software just a clever marketing illusion?