Founder/Developer/builder of the HTA POTS Sovereign Fabric that runs O(1) complexity and does what Quantum and Classical Supercomputers cant yet.
Alberta Canada
- Tweets 5,153
- Following 235
- Followers 228
- Likes 30
just wondering where i can get a refund for my pro subscription that was a complete waste of fking time as Astra is a piece of defective trash and you should be sued for false advertising if not producing defective garbage @DarioAmodei @OpenAI i hope you ipo soon so it craters!!
2
14
Been awhile since I’ve created a meme… I mean a DEADLY meme.. Wednesday it will land and will be one of my very best AND comes with a new word. 😱🙄🤡☢️☢️☢️☢️☢️☢️☢️☢️☢️☢️☢️☢️
@taylorswift13 @VancityReynolds @elonmusk @FallonTonight @nbcsnl @SabrinaAnnLynn @oliviarodrigo #meme
2
12
I wonder if this whole Cornell situation isn’t a deliberate distraction.. something similar happened awhile back and I would be curious was it around an election time?
2
25
well thats interesting now imagine AI running this.. hacking! just seems too easy. Reminds me of the early internet before firewalls wonder what a Tesla can do lol #canada #australia #china #usa tiktok.com/@abcnewsaus/video…
2
18
dont confuse this with quantum.. its me having fun developing a new memory primitive lol
🚨Benchmarked - The memory follows established distributed superposition memory theory almost exactly—then changes the failure mode: repair or refuse instead of silently returning bad data. 0 wrong-served outcomes across 192.24M read attempts.🚀🚀#ai #memory #computing #tech
1
13
how to make robots behave lol @mcuban
🚨Benchmarked - The memory follows established distributed superposition memory theory almost exactly—then changes the failure mode: repair or refuse instead of silently returning bad data. 0 wrong-served outcomes across 192.24M read attempts.🚀🚀#ai #memory #computing #tech
5
🚨Benchmarked - The memory follows established distributed superposition memory theory almost exactly—then changes the failure mode: repair or refuse instead of silently returning bad data. 0 wrong-served outcomes across 192.24M read attempts.🚀🚀#ai #memory #computing #tech
1
265
some more testing in the am then tear apart and ensure it ran right @martin_casado
A day of AI work comes to an end after burning through my token budget. I replaced the small-model Phi reader with frontier Claude Fable and finally got the full bounded-memory pipeline working end to end on held-out English and an official LongMemEval pilot.
#ai #tech #canada
interesting when you read up on current literature and see who is working on similar ideas yet different approaches.. wasnt expecting @brainchip in there
#ai #tech #canada
the AI experiments continue...
“Demonstrated deterministic, model-independent semantic state held at exactly 13,312 bytes through one million structured propositions, with an explicit information-loss contract and distance-invariant downstream reasoning on retained information.”
27
the AI experiments continue...
“Demonstrated deterministic, model-independent semantic state held at exactly 13,312 bytes through one million structured propositions, with an explicit information-loss contract and distance-invariant downstream reasoning on retained information.”
325