Your brain on ChatGPT is a slacker & how the VC A.I. playbook has changed
This MIT paper is the first formal research of the effects of ChatGPT on the brain I’ve seen. Researchers hooked people up to brain-measuring EEG machines and had them write essays. The group using only their brains showed the most neural activity, a full cognitive workout. The group using Google Search still got a decent pump in. But the group using ChatGPT? Their brains basically laid their feet up and got to rest.
The study showed that crucial brain networks for planning and organizing ideas went quiet, literally outsourcing the hard work to the A.I. . They called this long-term effect "cognitive debt." The more you rely on the A.I., the less your brain bothers to fire up for the next task, creating a cycle of dependency. As a bonus, the study also found the A.I.-assisted essays were more generic and the writers felt less ownership over them.
So the bad news is that the convenient shortcut that feels like it's helping is actually preventing your brain from practicing what it’s supposed to do. It might be that just using Google is a far better mental exercise, because you still have to do the work of filtering and synthesizing.
VCs are changing their approach to A.I. startups. The old playbook of having a team from Google, telling a good A.I. story and showing various exponential charts is dead. A.I. is not that novel anymore, and investors are tired of funding what amount to "features" that could be made obsolete by the next GPT-5 update. Instead of just getting excited about the tech, VCs are now asking much more pointed questions, like what’s the proprietary data moat in the business, scrutinizing how costs of compute and model inference are taken into account, or looking not only at the addressable market expansion brought by A.I. but also how pricing will settle down in the future once initial competitive advantages in the first phases of the technology vanish.