That Dawkins bit is wild. Smart people might be the most vulnerable to this stuff because they trust their own read too much. If he can fall for it, anyone can.
Exactly. You can’t just rely on people’s appeal to their own authority anymore. AI is too convincing - people really do need to understand the mechanism to assess it with any level of clarity.
Great article! A question I'm pondering, in practice how do Forward deployed engineers work differently with a client to a Systems integrator (e.g. Accenture) implementing an Enterprise SaaS platform like Salesforce? Some of the recent announcements feel like re-hash of a well known model. Even hiring aggressively, Frontier labs will struggle to deploy large teams of people to large corporates like a McKinsey or an IBM would, this is very different business/way of working. They will need major partnerships with SI's if they really want to crack large Enterprise transformation. Microsoft have a major advantage here too.
Yeah, the models are quite similar. Where I think this is probably headed is that consulting firms increasingly work (maybe even contract in) with the labs, helping map pain points and enterprise workflows alongside engineers, so the engineers can build bespoke products or agents around them. Eventually, enough similar problems show up that some of that bespoke work becomes discrete products, with customisation around the edges.
The open question is whether the labs end up spreading themselves too thin across the whole enterprise backend, or whether they actually manage to own more of it. Interesting to see how this develops.
That Dawkins bit is wild. Smart people might be the most vulnerable to this stuff because they trust their own read too much. If he can fall for it, anyone can.
Exactly. You can’t just rely on people’s appeal to their own authority anymore. AI is too convincing - people really do need to understand the mechanism to assess it with any level of clarity.
Great article! A question I'm pondering, in practice how do Forward deployed engineers work differently with a client to a Systems integrator (e.g. Accenture) implementing an Enterprise SaaS platform like Salesforce? Some of the recent announcements feel like re-hash of a well known model. Even hiring aggressively, Frontier labs will struggle to deploy large teams of people to large corporates like a McKinsey or an IBM would, this is very different business/way of working. They will need major partnerships with SI's if they really want to crack large Enterprise transformation. Microsoft have a major advantage here too.
Yeah, the models are quite similar. Where I think this is probably headed is that consulting firms increasingly work (maybe even contract in) with the labs, helping map pain points and enterprise workflows alongside engineers, so the engineers can build bespoke products or agents around them. Eventually, enough similar problems show up that some of that bespoke work becomes discrete products, with customisation around the edges.
The open question is whether the labs end up spreading themselves too thin across the whole enterprise backend, or whether they actually manage to own more of it. Interesting to see how this develops.