In a recent blog post by machine learning researcher Nathan Lambert it’s made evident that the pressure experienced by the once-sleepy AI research industry is enormous in the wake of ChatGPT:
“I’m oscillating between the most motivated I’ve ever been and some of the closest to burnt-out I’ve ever felt. This whiplash effect is very exhausting.”
This is due to ChatGPT being regarded as the “iPhone moment” for AI/ML/etc. This technology used to just exist in research labs, universities, and the dusty corners of some large tech companies — but now OpenAI products are used by over 100 million people worldwide. Take a second to let that sink in.
Forget organizations for a moment as Nathan’s piece focuses mostly on the individual implications of the AI boom: the siren song of a quick cash grab during this monumental shift, the challenges of prioritizing virality over scientific substance, the maddening effects social media algorithms have on the researchers who both built them and are now subject to them — and finally the crushing ethical weight of building the potential to wipe out all of humanity. No big deal.
“All of these low-level concerns make working in AI feel like the candle that burns bright and short.”
Professionally, AI researchers are confronted with a major question: as billions of dollars are flying around to every AI and ML project that achieves viral public acclaim, do I keep my head down and continue my work, or do I pull the ripcord and cash in by starting a company like thousands of my peers are doing right this minute?
That doesn’t even address the issue of “scooping” in the AI community. With so much of the research being done open source and in public, labs are developing similar technologies on similar timelines. Usually, research like this is done in the spirit of open sharing to push science forward for the betterment of humanity — but now that massive financial incentives are attached to being “first” it’s just not as simple any longer.
“Prioritization is really hard these days. If you’re obsessed with being first, the goalposts will keep moving as the next models get released. The need to be better and different is really strong.”
ELSEWHERE ON THE INTERNETS
Replacing bad systems with bad systems
Electric grids fed by renewables need a different kind of plumbing
“Awesome” ChatGPT prompts (yes we checked, they’re novel)
Hubble sees possible runaway black hole creating a trail of stars
Alcohol recovery startups Monument and Tempest shared patient’s private data with advertisers
