A confident decision can still be wrong.
Full episode at daax.ai/podcast/epis...
A confident decision can still be wrong.
Full episode at daax.ai/podcast/epis...
Full episode at daax.ai/podcast/epis...
Full episode at daax.ai/podcast/epis...
In this clip from our TokenDrop podcast S1E28 Sunil Baliga discusses Jev and hallucinations.
See the full episode at daax.ai/podcast/epis...
In this clip from our TokenDrop podcast S1E28 Sunil Baliga discusses Jev and hallucinations.
See the full episode at daax.ai/podcast/epis...
In AI, we barely verify at all. Christian Szegedy, who co-founded xAI and former AI Researcher at Google: "You can say it's even reckless."
TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #AISafety
In AI, we barely verify at all. Christian Szegedy, who co-founded xAI and former AI Researcher at Google: "You can say it's even reckless."
TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #AISafety
Christian Szegedy, who co-founded xAi and former AI Researcher at Google, on why checking outputs already gets you a very long way.
TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #AISafety
Christian Szegedy, who co-founded xAi and former AI Researcher at Google, on why checking outputs already gets you a very long way.
TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #AISafety
How do you forbid a behavior you can't even name?
Christian Szegedy, who co-founded xAI and former AI Researcher at Google, on TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #AISafety
How do you forbid a behavior you can't even name?
Christian Szegedy, who co-founded xAI and former AI Researcher at Google, on TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #AISafety
Christian Szegedy, who co-founded xAI and former Google AI Researcher, on the core AI safety problem.
TokenDrop S1E27
daax.ai/podcast/epis...
#AI #AIAgents
Christian Szegedy, who co-founded xAI and former Google AI Researcher, on the core AI safety problem.
TokenDrop S1E27
daax.ai/podcast/epis...
#AI #AIAgents
Christian Szegedy (co-author of Inception and Batch Norm) on ego and engineering.
TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #Engineering
Christian Szegedy (co-author of Inception and Batch Norm) on ego and engineering.
TokenDrop S1E27.
daax.ai/podcast/epis...
#AI #Engineering
Christian Szegedy, who co-founded xAI and former AI Researcher at Google, on TokenDrop S1E27.
Full episode at daax.ai/podcast/epis...
#AI #FormalVerification
Christian Szegedy, who co-founded xAI and former AI Researcher at Google, on TokenDrop S1E27.
Full episode at daax.ai/podcast/epis...
#AI #FormalVerification
In this clip from TokenDrop S1E26, Nav Mathur explains why plausible answers aren't enough when AI is making enterprise decisions.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop S1E26, Nav Mathur explains why plausible answers aren't enough when AI is making enterprise decisions.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop S1E26, Sajjad Khazipura explains why limiting what an AI can say isn't the same as ensuring it makes the right decision.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop S1E26, Sajjad Khazipura explains why limiting what an AI can say isn't the same as ensuring it makes the right decision.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E26, Sajjad Khazipura of DaaX explains why deterministic outputs alone don't guarantee correct decisions.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E26, Sajjad Khazipura of DaaX explains why deterministic outputs alone don't guarantee correct decisions.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E26, Sam Pooni of DaaX explains why making AI more deterministic may be less about adding capabilities and more about removing paths the system should never take.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E26, Sam Pooni of DaaX explains why making AI more deterministic may be less about adding capabilities and more about removing paths the system should never take.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E26, Nav Mathur, Co-Founder of Cerebrix AI explains how small probabilities of error can compound dramatically across multiple reasoning steps.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E26, Nav Mathur, Co-Founder of Cerebrix AI explains how small probabilities of error can compound dramatically across multiple reasoning steps.
Full episode at daax.ai/podcast/epis...
In this clip from TokenDrop podcast S1E25, the pattern behind DaaX's own approach to multi-agent systems.
Full episode daax.ai/s/f8zud9
In this clip from TokenDrop podcast S1E25, the pattern behind DaaX's own approach to multi-agent systems.
Full episode daax.ai/s/f8zud9
He tells it to apply the fix, but he won't let it happen automatically. In this clip from TokenDrop podcast S1E25, Peter walks through why.
Full episode daax.ai/s/f8zud9
He tells it to apply the fix, but he won't let it happen automatically. In this clip from TokenDrop podcast S1E25, Peter walks through why.
Full episode daax.ai/s/f8zud9
In this clip from TokenDrop S1E25, Peter explains how a job that used to require a full engineering organization now runs on a fraction of the headcount.
Full episode at daax.ai/s/f8zud9
In this clip from TokenDrop S1E25, Peter explains how a job that used to require a full engineering organization now runs on a fraction of the headcount.
Full episode at daax.ai/s/f8zud9
Peter Suaris, co-founder of AxPro Semi, is direct about this: AI is very powerful in the hands of an expert. In this clip from TokenDrop S1E25, why the person prompting the AI matters as much as the AI itself.
Full episode at daax.ai/s/f8zud9
Peter Suaris, co-founder of AxPro Semi, is direct about this: AI is very powerful in the hands of an expert. In this clip from TokenDrop S1E25, why the person prompting the AI matters as much as the AI itself.
Full episode at daax.ai/s/f8zud9
The four-step loop behind how 10,000 AI agents converged on a verified proof in 88 hours. Clip from our TokenDrop podcast S1E24.
Full episode at daax.ai/s/l99eg5
#AI #NeuroSymbolic
The four-step loop behind how 10,000 AI agents converged on a verified proof in 88 hours. Clip from our TokenDrop podcast S1E24.
Full episode at daax.ai/s/l99eg5
#AI #NeuroSymbolic
Add a third body to a two-body gravity problem and every tiny move changes every force at once.
Clip from our TokenDrop podcast S1E24 on what finally cracked it.
Full episode at daax.ai/s/l99eg5
#AI #HPC
Add a third body to a two-body gravity problem and every tiny move changes every force at once.
Clip from our TokenDrop podcast S1E24 on what finally cracked it.
Full episode at daax.ai/s/l99eg5
#AI #HPC
Only together do they work "like magic." In this clip from our TokenDrop podcast S1E24, we explain why DaaX's own tagline — LLMs need a helping hand — is exactly what made OpenAI's Navier-Stokes result possible
Full episode at daax.ai/s/l99eg5
Only together do they work "like magic." In this clip from our TokenDrop podcast S1E24, we explain why DaaX's own tagline — LLMs need a helping hand — is exactly what made OpenAI's Navier-Stokes result possible
Full episode at daax.ai/s/l99eg5
Clip from our TokenDrop podcast S1E24 on OpenAI's Navier-Stokes result.
Watch full episode daax.ai/s/l99eg5
#AI #NeuroSymbolic
Clip from our TokenDrop podcast S1E24 on OpenAI's Navier-Stokes result.
Watch full episode daax.ai/s/l99eg5
#AI #NeuroSymbolic
The real innovation behind OpenAI's Navier-Stokes result wasn't scale. It's a verifier that hands back a counterexample instead of a flat rejection. In this TokenDrop S1E24 clip, we explain why that distinction matters
Full episode daax.ai/s/l99eg5
The real innovation behind OpenAI's Navier-Stokes result wasn't scale. It's a verifier that hands back a counterexample instead of a flat rejection. In this TokenDrop S1E24 clip, we explain why that distinction matters
Full episode daax.ai/s/l99eg5
OpenAI took 88 hours to disprove global regularity for Navier-Stokes using 10,000 AI agents.
In this clip from TokenDrop S1E24, we explain why intelligence, not just compute, is what turned 88 years into 88 hours.
Full episode at daax.ai/s/l99eg5
OpenAI took 88 hours to disprove global regularity for Navier-Stokes using 10,000 AI agents.
In this clip from TokenDrop S1E24, we explain why intelligence, not just compute, is what turned 88 years into 88 hours.
Full episode at daax.ai/s/l99eg5