Realization of the day:
I’m not actually driven by money. I want comfort and to provide for my family, but I don’t care about Ferraris or flexing.
What I’m really chasing is status through trust. Being widely known, being respected, having access to other respected people I can directly exchange value and collaborate with.
Being pro-AI requires a degree of humility that most people do not have, as well as high openness and tolerance for change/the unknown that ego inflation does not make possible
@ChrisMBSLatham@WisprFlow Okay fair. I wasn't really talking about the tool though. I was talking more about the act of speaking into your computer in an office setting.
Didn't know if that was commonplace yet.
Serious question for anyone that works in an office with other people:
Is there anyone in there that just rips @WisprFlow all day?
I can't even imagine working any differently now.
Do you just have a bunch of people in the office talking constantly? Or is everyone still clicking and typing? (yikes)
I feel like this is going to be a human social evolutionary problem.
This was obviously tongue-in-cheek. I'm very empathetic to their situation.
However, I also live in reality. No one has ever improved their situation by coping for an extended period of time.
I'm doing my part every day to try to educate anybody that wants to learn about this new world that we find ourselves in.
Adapt or die. Not my fault.
My new, honest template reply for when someone asks me for a free "pick my brain" meeting (after I've already provided so much info online for free).
I've struggled with this for a while because I don't want to sound like an asshole, but I also need to focus on providing for my family.
Steal it if you also have this problem.
"Honestly, I can't take 1-on-1 calls unless it's paid coaching or client work.
I'm heads down on the business and I have to put every hour I've got into work that pays.
Nothing personal at all, not trying to act like some big time guy or anything.
That said, I don't want to leave you hanging - here are some options:
1. Pretty much everything I'm learning and teaching goes in my free newsletter: https://t.co/2YBSwDUzyS
2. I show what I'm building and my thinking on YouTube: https://t.co/aGdiPK0hNN
3. All of my free resources and systems, ready to download and use: https://t.co/INQhle6lkM
4. If you're trying to get into building stuff for clients, that's exactly who my community is for: https://t.co/MKUcwa7uef
5. For 1-on-1 coaching/advisory calls, it's $299/hour (or 50% off for community members): https://t.co/ljFjCNs1GY "
@mattmedeiros Arguing against the model you are hawking is a bold move but I'm here for it
Outreach pipeline is done btw...
AND MY AIOS AUTOMATIONS HAVE MULTI-LLM SUPPORT NOW
Used Claude Code + browser control to submit trademark applications for Motive11 and Modern Builders Society yesterday.
Compare that to how much time it took me to do it manually for Fueled by Progress back in 2021...yeeesh
I love AI.
All Cloud AI is _completely_ down? You're right, we're both fucked on the thinking. Neither of us builds shit. Welcome back to 2021, cool. (We still both have local AI, but let's set that aside.)
But if you're moving the goalposts, back to my original point:
Aside from a total Cloud AI catastrophe, the difference is when my preferred LLM (Claude) goes down, I click a different button. Codex, a local model, whatever's up. Same environment, same files, same context, same everything, same automations. Nothing moves, nothing breaks.
Grok Bot goes down - you're pinging Elon, opening a different environment, your Grok Bot routines break, you contextual accuracy suffers, I could go on...
I win in this game of shit that doesn't happen that often anyway.
The only thing we should caring about is making a bunch of money so we could somehow afford our own compute.
So, if you'll excuse me, I have an outreach pipeline buildout I need to orchestrate.
Asked my clanker to settle this. Here's what it said:
--
"Mark's AIOS has 33 automations. 28 of them never call a model. They're plain code Claude wrote, and they run on schedule whether Claude is up or down. Only 5 use a model, for judgment calls.
Grok Bot routines work differently. Per xAI's own docs, a routine's instruction is "what the Bot does on each run," and "each run spends usage." The Bot is the model. So every run of every routine is an AI call. Grok goes down, all of them stop.
So it's not that Mark uses less AI. He probably uses more, but he uses it to build internal software. The difference is where the dependency sits:
Grok Bot: AI is the worker. Needed on every run.
Mark's AIOS: AI is the builder. Needed to make and change things, not to run most of them.
Second difference: AIOS isn't tied to one model. The code, the data, and the context live on Mark's machine and accounts. His editor already switches between Claude, Codex, and a local model.
The 5 automations that do use a model are being moved onto a swappable slot so they fall back on their own as well. That part isn't finished yet."
--
Moral of the story: Control your data, own your software, have several inference options.