#59 This is How We Lose Our Jobs
Exploring the the Who What and Why of AI
In this issue, we discuss AGENTS. The Mr. Smith kind. The kind that follow you around, learn all your skills, then quietly sell themselves on the Bot Market for a few TaskTokens™ per day.
But, this is also the exciting part. This is how we get to stop working. This is how the powers that rule the world lose their control over our every move. Why? Because we’ll be out in the garden with our organic squash and quacky little ducks that help us eat the bugs.
Okay, let’s dive in. 🚀
Agent.Ai - is a perfect example.
Dharmesh, is one of the founders of Hubspot. He’s a programmer. Oh, and he’s a billionaire.
He understands technology, programming, and what the world of agents is starting to look like. So, what does a curious billionaire do? He drops $125K on the best domain name, and start coding. It’s free. Sign up here and get 50 extra credits.
Who Will Be Affected by AI Agents?
Anybody who works on a computer now - a lot of us - and those who do tasks with their muscles and eyes and brains - those are the people who’s jobs are at risk. We’re building literal robots these days and the Agents, as mentioned above, will be able to think, reason, calculate, search, read, write, etc. etc.
Highly paid, white-collar work won’t disappear instantly. But the threat is imminent. The transition will move at the speed of trust. The more we trust a bot ( digital or physical ) to get a result we can live with, the less we’ll rely on humans. So, we’ll finally get to do the things we want to do - but have never had time for. We’ll do art, music, and spend time with our pets and nature. It’s actually going to be pretty cool.
How Do Agents Work?
Right now, agents need to be focused. The best results occur when we give them a single objective ( or at least one that’s limited in scope ) for them to work on. This allows them to go through an iteration cycle where they do a little bit of work - analyze their work - then make adjustments to their approach until they get the result their human master ( for now ) requires.
So, when you think agent, just think “single task” for now. But, we can have lots and lots of agents. They’re all just focused computer programs working with large language models right now. As the models get better, the output of their work will improve. As it improves, our trust in their ability will improve. When that happens, we’ll start hiring them.
Create Your Own Agents
The near future will have a lot of opportunities. Agencies will scrape together a few bots and automation scripts, call them agents, give them an AI generated avatar, and call them something like Steve or Serena. They’ll be your agent. They’ll go do research… extensive research that would take you days and weeks. And.. because it’s so convenient, and such a time saver, you’ll think - “Meh.. it’s probably worth it.” - and in that purchase, your trust level will increase. You won’t tell your competitors, of course, and they won’t tell you, but Steve’s owner’s marketing department will do their best to ensure you know about him and his amazing research skills and the speed — …”did we mention it’s 10x faster than a human?”
Summary
Agents are here. Sign up. Get a company report and watch the extent of research that’s possible in only a few minutes. Then think how long it might have taken a college intern to do the same level of work - if they even could.
Ai gone wrong😱
Thanks for reading! See you next week.
WhoWhatWhyAi is a reader-supported publication. To receive new posts and support our work, consider becoming a free or paid subscriber.