![](https://discuss.tchncs.de/pictrs/image/a99d6771-39df-45b5-9d4e-a213ed480f50.jpeg)
![](https://awful.systems/pictrs/image/8651f454-1f76-42f4-bb27-4c64b332f07a.png)
i think that openai also wanted to solve their problems with fusion, but they got a step further, they made a startup for this. not normal nuclear power plant hot rock machine, no, they want tech that is perpetually Just A Decade Away. it makes some perverse sense if your funding is dependent on misguided hype only
dunno about rockets, but antenna thingy works only because you can simulate performance of antenna very reliably, precisely and quickly. This data was fed back, random small changes were made, things that were an improvement passed to the next iteration. Not sure how this approach is called but none of it is LLM