Garry Tan on X: "https://t.co/BA5kVYqcji" / X<br>Post
Log inSign up
Post
Garry Tan
@garrytan
99.1% Totality<br>In 2024 we flew to rural Texas outside Austin to see the total eclipse and ran out of charge on the Tesla we rented at the last minute. The ranch we stayed at was already maybe 99.1% so we decided maybe that was enough.<br>Here is the thing nobody tells you about an eclipse: 99% is not 99% of the experience. It is zero percent of the experience.<br>At 99% the sky is a little dim, like someone turned down a dimmer switch in a room you are still working in. You squint and wonder what the fuss is about. At 100% the corona comes out, the temperature drops, birds go quiet, and grown adults start to cry. There is no partial credit. The last 1% is not a rounding error. It is the entire phenomenon.<br>And here is the part that took me a while to understand: almost-good and undeniably-good are not neighbors on a line. We talk about quality like it is a gradient you climb, where 99 sits just below 100, one more push away. It isn't. The boundary is the thing. Totality is not the top of the dimmer switch. It is a different state of matter. You do not get to it by turning the knob a little further. Something has to flip.<br>I have been thinking about this because I keep meeting founders who live at 99%, and I am learning they are the hardest ones to help.<br>Here is what changed. AI collapsed the cost of building, so the product slide now looks the same for everyone. In any given week, several different founders can walk in with what is functionally the same pitch — same category, same wedge, same screenshot. The deck is identical. The demo runs. The TAM math checks out. From the outside you genuinely cannot tell them apart.<br>But they are not the same, and the difference is not the idea. It is who can take the idea across the boundary into a real artifact at 100% totality. When you and the user and everyone around you can feel the transcendance.<br>I know many founders like this, and I want to be careful how I say it, because they are to a T, good, earnest, so close and I am rooting for every single one of them. Everything they make is one degree off. The first time experience is almost right. The positioning is almost right. The thing people will pay for is almost there. These founders take feedback well and ship the next version, and the next version is also one degree off. Maybe 0.9 degrees off.<br>He is not failing. He is the opposite of failing. He is close enough that you keep driving and pushing, telling yourself the next degree is coming.<br>And here is the gentlest, truest way I can name the actual problem: he can feel the totality. In his own head he is standing in the corona. He knows exactly how good the thing is supposed to be. What hasn't happened yet is the transmission.<br>The product doesn't carry the feeling across to the user.<br>The user gets the dim sky.<br>The conviction is real and it is all on his side of the glass. That is not a character flaw. It is the single hardest thing in building: making the thing you can see become the thing a stranger can feel. Most people who are one degree off are not lazy or dumb. They are carrying a corona nobody else can see yet.<br>The premium used to be on can you build it. Engineering was the bottleneck, so the rare skill was making the thing exist at all. AI took that bottleneck out. Now the thing exists for everyone, which means the rare skill moved one notch downstream, to can you get it across the line. Not almost across. Across. Good enough that the corona comes out for someone who isn't you.<br>You cannot price this from the outside. Almost-good and good photograph identically; the dim sky and the dark sky look like the same sky in a deck. You only learn the difference by watching someone ship under real conditions, with a real problem, and seeing whether the thing finishes or merely assists.<br>Which brings me to the strangest lesson, and the most useful one. You would think that in a world where AI does more of the building, who you stand next to matters less. It is the opposite. The danger of 99% is that from inside it, it feels like totality. The dim sky looks bright enough when your eyes have adjusted to it. The only reliable way to find out you are in twilight is to stand next to someone who has been in the dark: a builder whose work actually flips, whose artifacts go all the way across. Proximity to the real thing recalibrates your eyes. It kicks you out of the local maximum you didn't know you were sitting in, because suddenly you can see the gap between your 99 and their 100, and you can't unsee it.<br>So the advice inverts. Being around spiky, undeniable builders is worth more now, not less. Not for the tips. For the calibration. They are the ones who show you where the path of totality actually runs, so you stop parking one county over and calling it close.<br>We did not see the eclipse in Texas. We saw 99.1% of it, which is to say we saw a slightly dim afternoon. I think about that...