How TypeSafe AI came out of stealth with Jev on X
40.1M views, ranked 16 of the 2367 launches we track. Doomers AI ran the launch; it reached 26,762 times the account's following.
2:56
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x
In the Doomers AI Launch Index
A Launch Score of 92 for September 2026, ahead of 100% of that month’s launches past 100K views. See the Launch Index and how the score works.
How it compares
A launch means nothing against no one. These are the launches it is actually comparable to, and where this one sits among them.
| Compared with | Launches | Median views | This one | Rank |
|---|---|---|---|---|
| Frontier labs launches from accounts under 5K followers | 33 | 428K | 93.9x | 1 |
| Product news the same kind of announcement | 881 | 722K | 55.6x | 6 |
| under 5K followers accounts of a similar size | 775 | 208K | 193.2x | 1 |
| Every launch all of them, for scale | 2367 | 595K | 67.5x | 16 |
The read
TypeSafe AI came out of two years of stealth on 15 September 2026 with Jev, a new frontier model, and the announcement went out from the founder's own account rather than the company's. Read the ratios rather than the totals; the post was hours old when this was written.
The post led with a credential a reader could verify in one line: the founder co-created RLHF and ChatGPT at OpenAI. Then it stated the question the company exists to answer, then the two years in stealth, then the product, then three checkable claims about speed and cost.
Nothing in the first sentence asks the reader to care about a company they have never heard of. It asks them whether the person who built the thing they use every day has a point.
The video runs just under three minutes, which is long by the received wisdom and normal for the launches in this directory that travel furthest.
Launch videos over two minutes do twice the median of videos under thirty seconds. A demo that shows the model doing the whole job holds attention, and a ranker that grades on watch time rewards it.
What separates this launch from most of the large ones we hold is what happened after the views. Saves ran at several times the directory median and reposts well above it, on a post that also reached more than a hundred times the following of the account that sent it.
Most launches at this size are carried by the ranker and endorsed by almost nobody. This one was carried and endorsed, which is the pattern that turns into signups, and it is the number we grade a launch on when there is a considered purchase behind it.
Doomers AI ran the launch strategy and the distribution. The announcement was seeded to between 80 and 100 specific accounts, engineers, founders, builders and creators whose audiences overlap the buyer, so the right people were reading it in the first hour.
Twelve days later, Sequoia partner Pat Grady used the launch as an example of how fast the market moves: Jev went from zero to $100 million in revenue in its first seven days, and the case study has the source.
The pickup beyond that was organic. What that involves is on the services page, and if you are launching something, tell us what it is.
The fuller account, with the two accounts' growth, the seeding, and the video credit, is the TypeSafe AI case study.
Questions about this launch
When did TypeSafe AI launch?
TypeSafe AI posted the launch on X on 15 September 2026.
How many views did the TypeSafe AI launch post get?
40,142,284 views at the last read, ranked 16 of the 2367 AI launches Doomers AI tracks, 26,762 times the account's following.
Who posted the TypeSafe AI launch?
@CompleteSkeptic, which had 1,500 followers when it posted.
Planning something like this?
We run launches on X for AI companies, with creators who already hold the audience you want reaching.