Sankaet Pathak
speaker
95 appearances
1 recordings
1 series
first heard Jul 2026
last heard 31 Jul
Sankaet Pathak’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.
Appearances
But in net, I do think it is going to increase our quality of life and will lead to the age of abundance.
Well, Piers, just to be fair, I am still quite skeptical of OpenAI's characterization of what happened.
So I would just give it some more time and figure out what exactly the root cause was.
A lot of the scenarios, I personally firsthand have information on prior instances like this, where Anthropic has also claimed that the AI somehow blackmailed some employee at the company.
And they're massively exaggerated and overblown.
And in any real world scenario, they pose no threat.
So I still do not know exactly what happened in the open AI scenario.
But then to your question of self-replication, I do think there are significant dangers where AI kind of gets out of hand is one, when it can make itself better with less compute.
Because today we know on our fingertips by and large, all of the big compute clusters.
So if the risk is outsized,
to the reward, we will shut training down for almost all of the AIs.
As soon as that changes, which is AI can self-replicate, self-improve with less compute, and has a electric footprint that is significantly less to what it has today,
That is for sure an unbounded risk for humanity, and that is something we have to take very seriously.
As it stands today, however, they need massive compute clusters to get reasonably better over time.
And almost every single scenario where people have tried to portray that their AI is somehow so good and so crazy have all been overblown.
Our entire company still uses flawed code and it still makes tons of mistakes.
So there's no scenario in which it's remotely close to being super intelligent.
And I'd be highly skeptical of what OpenAI is saying.
My very skeptical mind says everything they're saying, they're saying just to be able to get a psychological edge on Anthropic because they're falling behind.
I still don't know exactly what happened there, but historically all of those charts are extremely overblocked.
Showing 21–40 of 95 · page 2 of 5
← Previous
Next →