Greg Isenberg’s viral thread on “20 Mega Shifts” that make 2026 the best time to build a startup captures something real about this moment. But a few points deserve some pushback or refinement.
On Social Distribution
The point about “FYP-ified” social is about not needing to grow a following first. You can write a couple quality posts and get seen way more than the old follower-only view allowed. That’s genuinely democratizing for creators starting from zero. The smart play is converting that rented algorithmic reach into owned channels—email, SMS, community. Creator Governed Content is needed to start to put the power back in the hands of the creator.
On SaaS “Imploding”
Maybe this should have said traditional SaaS is imploding. AKA large companies with hundreds or thousands of engineers that only offer a simple product.
Simple products will be easy to replace, but many people will have “real world fears.” A plumbing company isn’t experts in taxes or accounting and it will be a while before they trust the AI to just get it right. If they mess it up, they’re in legal trouble. So they’ll pay to ensure they have software that is correct. Maybe the software comes with accuracy insurance. That liability shield is a moat.
There’s also deep domain knowledge in some areas. Software that just pipes data around or creates dashboards is probably done, but something giving medical recommendations or reviewing compliance is different. You may want an expert in that subject to have built the tool. People will be willing to pay for that expertise and trust.
On Moats
Greg asks “what’s your moat?” and suggests distribution, customer service, brand, data. AI will commoditize more than just software eventually, but that’s probably a bit farther out. Software agents are already in production and working at full steam. But at some level there’s a deep connection that makes humans trust and prefer other humans.
The person who likes to go to their hairdresser once a week or once a month is very unlikely to prefer getting their hair done by a robot. Period. Some people will because it’s cheaper, but there’s an innate desire from lots of people: “my human hairdresser.” That person is more valuable than the cheaper robot haircut, even if the result is the same. There will always be people who dramatically prefer and support a person over AI tooling, or a faceless company, or even someone who has a face digitally but they can’t meet in person. The brand being the individual person has value.
The data piece is also an understated moat. AI needs a lot of data or context to get the most optimal answer. While a lot of data can be searched up online, which provides zero moat, private conversations, private meetings with business partners or coworkers is information that can make or break AI tooling. Capturing that, having that context, is not something you can replicate with software alone. Those interactions have to happen. There’s real value there.
On Funding and Creator Economics
The MVP (0.1) and even 0.2 maybe no longer require funding like they used to. 1-2 people can handle this rather than the 10-100 that were needed before. You might even get to 2.0 before needing any cash infusion for scaling, particularly if your operational costs are being met by your product. If you’re cash flow negative purely on an operational basis (excluding headcount as SG&A), then having low or high headcount doesn’t matter. You need investment regardless.
When you go to scale from MVP to a 1.0 product, that’s where you can leverage creators. You can pay creators in things other than money, like equity in the product, or through ideological alignment. Showing them that this product is going to change the world in a direction they want it to go. Start using my product and start advocating for it.
The cost of acquisition goes down significantly in the creator/community model compared to blanket advertising. And the lifetime value is higher because someone who joins through community trust churns at a fraction of the rate of someone who clicked an ad.