In short
We did not start with chips. We started with a question about why organisations cannot see themselves, and we went looking for the place where that blindness costs the most. It is a silicon program: the longest chain of commitments, across the most companies, with the least forgiving date.
Five founders, three continents, one question
Tymeline was founded by five of us, deliberately spread across three continents from the first day: engineering in Bangalore, strategy and finance in California, headquarters in Austin. The question we started with came out of the research we published in 2024 on what we called Autonomous Organizational Intelligence: why does an organisation with more data about itself than ever still find out what happened at a meeting?
A framework is only worth something when it runs a real program. So we went looking for the program where the question hurt most.
Why silicon is the hardest program there is
A chip program is a two-year chain of commitments across a dozen tools and several companies: the design team, the IP vendors, the foundry, the packaging house, the firmware team, the customer waiting for samples. Each link lives in a different system under a different manager. The date at the end does not move for any of them, and the cost of missing it is a mask set, a quarter of the whole team, and often the socket.
Industry data says first-silicon success is at a twenty-year low and most chip projects run late. When you read the post-mortems, the engineering is rarely the cause. The cause is a decision made months earlier that nobody could see the cost of. That is exactly the blindness our research was about, in the one industry where it is most expensive.
What we decided a program needs
Four things, and we named the product after them. A record: every commitment, its owner, its date, read from the tools rather than typed into slides. An alarm: the drift, the hour it starts, with what it breaks downstream. Hands: AI Employees that work out the fix and carry it out inside limits a person sets. And memory: what fixed it last time, kept for the next program. Underneath them, Kendrel, the brain that turns a thousand signals into one decision a VP can act on, and Tymeline ID, the identity and the signed record that make it trustworthy.
None of it is specific to silicon in principle. All of it was shaped by silicon in practice, because silicon does not forgive a vague answer.
What I believe this becomes
I think the next decade of engineering will be decided less by who has the best engineers and more by who can see their whole program at once and act on it the same hour. Chips are where that is being proven first, because the date is the hardest. The rest of engineering has the same disease and will want the same cure. We are building for silicon first because it is the proof, not the limit.