categories

HOT TOPICS

1Mby1M Virtual Accelerator AI Investor Forum: With Daniel Cohen, General Partner at Viola Ventures (Part 2)

Posted on Saturday, Nov 23rd 2024

Sramana Mitra: Which ones do you categorize as the ones that are going to be the harder to deal with?

Daniel Cohen: There are areas in the world that technology has disrupted and yet not disrupted at all. If you look at Healthcare, on one hand, of course our life’s changed completely, but then there’re some things that are exactly the same.

In some other tough areas, like construction or airlines or travel, there’re many changes yet the flight from London to New York is exactly the same as what it was 50 years ago. Probably business class is a little bit nicer, but it’s basically the same.

So the big question is what kind of areas or verticals – aerospace, travel, or healthcare- can be changed? Can you really change fundamentally some of the physics of those operations and can AI make a huge difference? The way to figure that out is just going take a long, long time and the changes will just take 10-15 years to come versus media creation, which is here today.

Sramana Mitra: Healthcare is obviously the biggest prize for AI. It has the biggest potential. In my opinion, it’s the most important vertical where AI can have an enormous positive impact, but it is also one where hallucination cannot be tolerated.

Hallucination can be tolerated in media creation, but it cannot be tolerated in healthcare.

Daniel Cohen: Of course, you’re right. Right now, our AI expectations are way above what AI can deliver. Although we’re willing to suffer hallucinations in some areas because the damage is not there, I think from a user perspective, our expectation is that AI can do at a high level, but there’s a gap there in what it can do.

Specifically on healthcare, there are things that we will see very soon, especially on research side and some things that are kind of more in its backside. But I think for diagnosis where we’re interacting with people, I think it’s going to take a lot of time before we can talk to an AI doctor.

Sramana Mitra: On that though, there are several gating items in healthcare. I’m not saying that there are, I will get to the multiple gating items. But if one of the gating items is hallucination, that can be controlled with small language models. You know, you don’t have to use these unwieldy large language models and let that lose on healthcare. Instead, if you control the language models, you can control hallucination.

Daniel Cohen: But do you get the same quality out of it? Does it match the quality of expectation of the user? Do people get the same value out of it that they want?

Sramana Mitra: Well, I think that there are other gating items there that may not get us to where we want to get to, which is the entrenched systems that need to be upgraded somehow. These are old architecture systems. In America, for instance, Epic dominates. And that product was developed a long time ago on a very, very old architecture. And for that to catch up to where we are right now in terms of AI potential and all that, that is definitely a gating item.

This segment is part 2 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: With Daniel Cohen, General Partner at Viola Ventures
1 2 3 4 5 6

Hacker News
() Comments

Featured Videos