Hi, I am looking for a founder with a good business idea that has to do with the enterprise AI space. I myself am a founder turned investor that invests on the Pre-Seed/Seed side. Feel free to DM or send me an email :)
Ok this has been bugging me for a few weeks. Everyone keeps saying AI-native companies are beating incumbents 10x their size because they've got better models. I don't buy it, most of these startups are running on the same base models as the incumbents.
What they actually have is a workflow with no legacy. Incumbents built their process around a person doing the work, and now they're bolting AI on somewhere in the middle as an assistant. That's a patch, not a redesign. A startup with a blank slate can flip it, software acts first, a human only gets pulled in on the edge cases it can't handle.
Where I'd push back on myself: this only works when being wrong occasionally is cheap. I don't think "software executes, human checks the exception" survives anything regulated or high-stakes. So it's not some universal unlock, it's an advantage in a specific band of use cases, and I think a lot of the hype just skips over which band that is.
(Talked this through with a couple people at G2C Ventures who invest in this stage of AI/enterprise stuff, they see the same split in their portfolio, some companies can run software-first and some very much can't.)
Where do you think that line actually sits? What workflows do you think "software-first, human-supervises" genuinely works for right now vs. where it's still just a pitch deck line?
A founder in our portfolio called a few months back with his best quarter ever. What stuck with me wasn't the numbers, it was how he opened the call. First thing out of his mouth: "You've gotten bad news from me for three years straight, wanted you to hear the good stuff first for once."
Got me thinking back over a dozen board meetings with him. He's never once tried to bury a miss. No "headwinds" language, no burying the ugly slide on page 14 of a 30-page deck hoping nobody asks.
Which made me sit with something I've been chewing on: I don't think most founders can afford to run their boards this way, even though I'd like to believe otherwise. That kind of honesty only works if you survive the one quarter where the truth is genuinely bad and your board doesn't panic. Most boards say they want transparency. Fewer actually reward it in the room when it costs them something.
So founders learn the lesson that hedging is safer than honesty. Then everyone acts shocked later when a company falls apart and the board had no idea how bad things had gotten.
That's why "just be more transparent with your board" advice rubs me wrong. It treats this like a founder discipline problem. I think it's an investor behavior problem. Punish bad news when it shows up, and you'll get less of it.
Curious if anyone's tested this from either side — founder or investor. Did honesty ever genuinely cost you?
This Jensen Huang clip is going around my feed again and it's bugging me enough that I have to write it out somewhere.
His argument is basically: AI automates tasks, not jobs, and radiology proves it. Demand for radiologists actually went up even after AI got good at reading scans, because the job was never "read the scan," it was diagnosis, judgment calls, talking to a scared patient. He then says the same logic carries over to software engineers and that's where I start to disagree with him.
Radiologists have a gate around them. Med school, residency, a license you can't just skip. Nobody can flood the market with radiologists even if there were suddenly infinite demand for them, so any freed-up capacity has nowhere to go but... more radiology, more careful diagnosis, whatever. Engineers don't have that gate. If AI makes one engineer 3x more productive, nobody is required to go find bigger ambitions for that extra capacity.
As a seed/early stage fund I'm on the investing side of some of these companies and I'm watching it happen right now. Smaller eng teams, same or better output, and it gets presented in board meetings as a genuinely good thing, nobody's treating it like something to be nervous about.
So, has anyone actually seen anything hold that line for engineers the way licensing does for doctors? Certifications, unions, anything?
A founder in our portfolio called a few months back with his best quarter ever. What stuck with me wasn't the numbers, it was how he opened the call. First thing out of his mouth: "You've gotten bad news from me for three years straight, wanted you to hear the good stuff first for once."
Got me thinking back over a dozen board meetings with him. He's never once tried to bury a miss. No "headwinds" language, no burying the ugly slide on page 14 of a 30-page deck hoping nobody asks.
Which made me sit with something I've been chewing on: I don't think most founders can afford to run their boards this way, even though I'd like to believe otherwise. That kind of honesty only works if you survive the one quarter where the truth is genuinely bad and your board doesn't panic. Most boards say they want transparency. Fewer actually reward it in the room when it costs them something.
So founders learn the lesson that hedging is safer than honesty. Then everyone acts shocked later when a company falls apart and the board had no idea how bad things had gotten.
That's why "just be more transparent with your board" advice rubs me wrong. It treats this like a founder discipline problem. I think it's an investor behavior problem. Punish bad news when it shows up, and you'll get less of it.
Curious if anyone's tested this from either side — founder or investor. Did honesty ever genuinely cost you?