Most teams I run into aren't new to AI, they're already running ChatGPT or Claude. The problem is they're using it in a manual way. Ask questions, get responses, copy and paste things between other systems.
What they need to graduate to level 2 is to onboard an agent. Give it a job and the permissions needed to finish it. Read the inbox, sort the routine stuff, update systems, and hand judgment calls to a person when needed. Humans stop being the bottleneck, and start having a real digital assistant that helps with their day.
So where do you start if you want to do this?
Before implementing a tool, document the process. Start with one bounded task. Inbox triage, invoice follow-up, client onboarding. Define the inputs, the outcomes, and whether a human checkpoint is needed for anything consequential.
Brick by brick, a lean team can take on more work without adding head count!
Interesting point on economic indicators, and I think this has been an underrated AI benefit.
I am the farthest person from a handy man, and using Claude + photos I’ve been able to fix my own gutters, troubleshoot minor appliance issues, and figure out why my central air wasn’t turning on.
Curious how many other people have found it enabling for fixing physical things?
I just fixed my dishwasher with the help of ChatGPT. A trivial task. I had been about to order a new one. So this software has increased the country's real wealth yet decreased the measured GDP. The main economic indicator is structurally incapable of registering the thing that actually makes people better off, namely the growth of knowledge
@cbwritescopy Remember hearing that copywriting GOAT Dan Kennedy only would allow people to contact him by fax. I understand that more and more every day…
@emollick I think everything should just be a work/cowork chat. Having the two separate contexts that can’t reference each other makes for a mess, especially with the nightmare of migrating a personal plan to teams.
@danielcberk I like the Lindy model of betting: only between friends and the stakes being something that strengthens the social bond like covering the bill/next round.
I recognize there are some bearish on AI arguments worth considering.
However, I would humbly ask that you try out a SOTA model instead of pretending things haven’t advanced since you tried GPT-4 years ago…
Fair trade, the meltdowns I’ve seen over the access roll out are a bit ridiculous. Remember folks, it costs nothing to be classy and show a little grace!
We will give one banked reset for every day you don't have access to Astra on your paid ChatGPT plan, starting today. Team is moving mountains to give access as fast as we can.
First one will land in ~ 3 hours. There is still time to create your account if you don't have one.
Always Be Following Up
Case in point: Today I signed up for something where I initially expressed interest to the rep over a year ago.
For a while the timing wasn't right, then for the last two months it honestly just wasn't making my list of things to do even though I was ready to move forward.
Never assume a "no" until you get one!
@GergelyOrosz Mostly agree, however I think the goal posts for our expectations are moving (which is a good sign!)
Two years ago this would have been mind blowing. I think the solo dev AAA experience game is closer than we think.
One for parenting: I love finding cool town rec programs for my son.
The issue: They are all behind portals that are separate for each town and not retrievable with http requests. You need to be driving a browser in the portal looking for what is available and setting age group filters.
I used Grok bot to dig through everything within my county that is in my son’s age range. Would have taken me an hour to do this manually, and now I can set it to be recurring!