What's up with all of these crazy technical developments after a period of relative stagnation?
Aside from it being merely random or selection bias, it's plausible that there's some kind of AI assist beyond what we know happening behind the scenes.
These baseless speculations about ChatGPT are somewhat ironic given that the article itself alludes to what has unlocked this specific advance:
"The RDRE achieved its primary test objective by demonstrating that its hardware – made from novel additive manufacturing, or 3D printing, designs and processes – could operate for long durations while withstanding the extreme heat and pressure environments generated by detonations."
It's much easier to imagine that steady advances in manufacturing, computational modeling, and "mundane" technologies like that are contributing to progress like what's described in this article... Rather than some super secret AI assistant.
Please, folks, don't give up on reading just because computers can do it too now. :)
People have been working on rotating detonation engines, and making steady progress, for years. The idea dates to the '60s, but i think work really picked up about a decade ago.
OH MY.... Came here to say "Wow, this is so much more exciting than the 123th LLM/GPT and the 2746th "this is my experience / use case with AI" (sorry) "but sure, some one will come and claim that: GPT-8 would have designed that alone given the right prompt!!"
...
and then need to read this. :(
I said a few days ago that it is very likely 3 letter agencies and other organisations have had GPT4 equivalent models for some time now. It doesn’t seem that unlikely that some branches of the US gov would have much more advanced models.
Why do people always say this? It's almost like a thought-terminating cliche. Thousands of companies competing with each other were not able to do it earlier, but the NSA always has better everything, by definition.
Maybe ChatGPT is helping cut through bureaucratic tape. I know it’s helped me become much more persuasive in my communication. If your goal is to convince the US government to spend money testing your rotational detonation engine, your odds of success go up if an LLM helps you write the email.
Aside from it being merely random or selection bias, it's plausible that there's some kind of AI assist beyond what we know happening behind the scenes.