Daniel Gross: prediction

Transcript context

…Hey, I’m Josh. Thank you for speaking. This is an awesome conversation. One big topic, in SF especially, is on automation. Are there any parts of your talent process that you think could be automated? Obviously, ATS is a thing, and with the rise of automation, maybe it’s pretty industry-specific. But are there any changes in how people seek talent that will come as things get more and more automated? I run a company that principally has — for our little corner of talent, meaning venture — tried everything under the sun in order to automate it. Like many processes, you basically split things into two. There’s the spam filtering process of basically weeding out people that don’t make any sense for us, for whatever reason — what they’re working on is noneconomic, or they don’t have the qualifications. That you can probably do with software. There’s a second step of it. Basically, okay, imagine this is Gmail. You’ve got rid of the spam, but now you’ve got to pick what in the inbox is important. That’s a much harder task. I’m sure it can be done in software, but I think it’s a bit more nuanced. The really tricky thing — in venture, in particular — is regressing on success is pretty hard, not just because the data points are pretty sparse, like, what great founders look like — maybe you have 10,000, which is not that helpful. Machine learning skill. But also, because everything changes all the time. The psychometric makeup of a great founder in, say, 2015, SaaS era, is someone like Fred Luddy, who started ServiceNow, who’s basically a sales machine, started a sales empire. It’s very different from who’s going to be a very good founder, say, working on transformer models, who’s going to be much more like Woz [Steve Wozniak] than Steve [Jobs]. Everything is shifting constantly, so it’s tricky. I assume a lot of automation can be done for that first step. I think, for the second step, you could. But the final thing I’d say is, in venture, in particular, you are rewarded so aggressively for making the right calls that you will be able to always afford the salary for people to review it. You’re penalized, of course, very aggressively by errors of omission, not commission. I think you’re going to always end up with an economic model where you can have people. This is obviously very different if you’re like McDonald’s and whatnot, and you’re trying to figure out, “Okay, who’s going to be able to flip burgers a year in?” But I don’t know enough about that field to apply it there. Thank you for attending. Thank you, Tyler. Thank you, Daniel. This was fun.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.