How to Choose Your Personal AI Agent

Nathaniel Whittemore (0:00)

.,We We are are officially officially drowning drowning in in personal personal agents agents. Between Between Muse Muse and and GrokBot GrokBot, and and OpenClaw OpenClaw, and and Hermes Hermes, and and DotsNow Dotsnow, and and Instinct Instinct, and and whatever whatever Anthropic Anthropic inevitably inevitably launches launches. this. This is is a a form form factor factor that that is is absolutely absolutely everywhere everywhere. And And given given how how much much context context, and and setup setup, and and tool tool access access, and and account account access access is is going going to to be be required required to to get get the the most most out out of of these these personal personal agents agents, the the cost cost of of switching switching could could be be kind kind of of high high. Given Given all all that that, today, today we we are are looking looking at at a a guide guide to to how how to to choose choose a a personal personal agent agent based based on on a a set set of of different different criteria criteria that that should should help help you you hone hone in in on on which which one one is is best best for for you you.

The The AI AI Daily Daily Brief Brief is is a a daily daily podcast podcast and and video video about about the the most most important important news news and and discussions discussions in in AI AI.

Today Today, we we are are talking talking about about personal personal agents agents. And And here's here's my my shtick shtick on on this this. I I think think it's it's early early enough enough that that having having some some amount amount of of skepticism skepticism that that this this particular particular form form factor factor that that everyone everyone is is racing racing to to implement implement for for you you will will ultimately ultimately be be where where things things land land.

However, the. The idea idea that that you you are are likely likely to to have have at at least least one one highly highly connected connected agent agent integrated integrated with with your your email email accounts accounts, your, your Slack Slack or or Teams Teams, and, and even even potentially potentially having having access access to to things things like like your your financial financial accounts accounts. is. Is going going to to be be increasingly increasingly normal normal. My My argument argument then then. is. Is that that even even if if you're you're not not sure sure that that this this is is exactly exactly a a fit fit for for you you, I I think think it's it's worth worth carving carving out out some some experimentation experimentation time time to to see see if if and and how how using using a a personal personal agent agent impacts impacts anything anything in in your your professional professional or or personal personal life life. Now Now, that, that does does not not mean mean you you have have to to give give it it access access to to everything everything to to get get a a real real sense sense of of it it, but but it it does does mean mean putting putting in in some some actual actual time time and and reps reps.

But your. Your time time is is precious precious. and. And there there is is too too little little of of it it. So So which which personal personal agent agent are are you you going going to to choose choose to to experiment experiment with with? Today's Today's show show is is all all about about that that question question, but, but we're we're gonna going to divide divide it it into into three three parts parts. I'll I'll go go through through a a high-level high-level framework framework for for thinking thinking about about it it, share share a a little little interactive interactive quiz quiz that that I I built built, that that you you can can use use yourself yourself after after you you listed listed this this episode episode. But But before before we we do do that that, I I want want to to read read some some excerpts excerpts from from Professor Professor Ethan Ethan Molek's Molek's most most recent recent post post on on his his One One Useful Useful Thing Thing blog blog, which, which is is called called The The Dot Dot and and the the Swarm Swarm. It's It's a a meditation meditation on on this this personal personal agent agent form form factor factor, and and why why even even someone someone who who watches watches things things as as closely closely as as Ethan Ethan does does, can can miss miss where where AI AI is is headed headed. Ethan Ethan writes writes. I I generally generally think think I I have have done done a a good good job job anticipating anticipating the the direction direction and and pace pace of of AI AI over over the the few few years years I've I have been been writing writing this this substack substack. But But I I think think I I recently recently got got something something fairly fairly large large wrong wrong. In In the the last last year year, I I have have been been posting posting about about how how I I suspected suspected that that humans humans would would have have to to approach approach working working with with agents agents as as a a manager manager, deciding, deciding how how to to delegate delegate work work to to agents agents and, and specifying specifying how how those those agents agents should should be be organized organized. I I thought thought that that getting getting agents agents to to work work effectively effectively as as a a group group would. Would take take careful careful construction construction akin, akin to to building building a a company company. and. And that that this this would would take take time time to to figure figure out out.