The Fast Iteration Cycle: How Progress Really Happens | Josef Fleischmann | TEDxTUM

Josef Fleischmann opens with a dramatic anecdote about witnessing the launch and explosion of a rocket his team developed over 7.5 years at a launch site in the Arctic. Rather than viewing this as failure, he frames it as a success because it provided critical learning data. He introduces himself as CTO and co-founder of ISA Aerospace and outlines his talk's focus: applying fast iteration cycles across various fields including aerospace, startups, and education.

Fleischmann contrasts traditional aerospace development with modern approaches. The old space industry model, dominated by NASA, ESA, and corporations like Boeing and Airbus, was characterized by years or decades of development, massive cost overruns, and risk-averse cultures. He cites NASA's Space Launch System as an example: supposed to launch in 2016 but didn't until 2022, with costs reaching $2 billion per launch. This conservatism ironically creates risk by slowing progress and increasing costs.

The fast iteration cycle concept, by contrast, involves bringing products to testing as quickly as possible, learning from both successes and failures, gathering data, and rapidly reiterating. The core principle is maximizing learning speed rather than perfecting designs before testing. Fleischmann provides concrete examples from ISA Aerospace: their fuel tanks, which comprise 75% of the rocket's external length and 90% of its mass at liftoff, were developed in just 2 years with extensive testing. These tanks must withstand 400 bars of pressure while holding liquid oxygen at -200°C while remaining lightweight and cost-effective.

Another critical example is their Akila rocket engine, developed in 4 years—remarkably quick compared to the Ariana 6's upper stage engine, which took over 28 years total development (starting in 1996). Fleischmann attributes this acceleration to 3D printing manufacturing, in-house testing facilities, and rapid iteration cycles. He emphasizes that rocket engines are particularly challenging because knowledge isn't documented; companies keep secrets, making iteration essential.

Fleischmann broadens the applicability of fast iteration cycles beyond aerospace to startups, student projects, and general professional work. He argues that successful students are those who iterate quickly by seeking feedback and adapting their learning to practice. He identifies key learnings from their development approach: avoid recklessness (calculate risks), extract learning from every test, push boundaries incrementally with each test, and apply products in operational environments as early as possible to gain real-world feedback.

Returning to the rocket explosion, Fleischmann reveals the team's response: despite initial shock, they immediately gathered all data, recovered hardware, iterated on both hardware and software, and rebuilt the rocket for another launch attempt. He reframes the explosion as a successful learning exercise rather than a failure, emphasizing how systematic response to setbacks drives success. He concludes by noting he didn't invent this principle at ISA Aerospace but has applied it throughout his life as a student, startup founder, and engineer, recommending his audience adopt it across their own endeavors.