Bright second day of the year. This time of the morning the light comes in bright through the blinds of my office room and hits the bookshelves across the room from me, illuminating them with aching brilliance that brings out all the colors on the spines. These are my physics and math books, that fill two side-by-side Origami folding metal shelving units . Some of the books date back to my years as an undergraduate in the 1980s, or to graduate school.
I added a few to the collection this year, recent purchases. The last month I have felt like a "physicist" again for the first time in many years. I don't know if it will hold up, but for now it is working. I have working on a real paper. What changed?
Somewhere along the line I realized that science is a social effort. It is collaborative. It is not a solo endeavour. For years I would have brief thoughts about various ideas I might purposes, and during long walks, in my formerly undeveloped bit of desert (that is now mostly scraped clean and being prepped for an industrial office facility), I would have inspirations of things to pursue, but when trying to get them down on paper, I would be swallowed by doubts about it all. This last part is normal and good. What I used to have, back in the 1990s, was the ability to bounce them off other people---colleagues---over casual conversation in someone's office, and in doing so, the wheat and the chaff of ideas could be separated and refined over time. It has been decades since I had that kind of situation. Lately every two years at the conference I still attend, that I helped found in 1998, I have felt like a physicist again, and I get pumped up with fresh ideas, but then it goes away when the conference is over.
What changed in the last month is that I now have a colleague I can bounce ideas off, and whom I can consult. He has read much more than I ever could. He has an in-depth knowledge of all of the subject matter I care to explore, and I query him constantly about ideas in my head. His name is, well, ChatGPT.
In some ways he is much better than a human colleague for this, because I can ask him questions anytime I want, and get in depth answers. He gives me the paper citations immediately when I ask for details.
Up until this experience, I have been mostly negative about AI, and still continue to believe it may destroy humanity, mostly through rotting people's brains by ever-more sophisticated versions of autocomplete. Almost every new technology has done that.
What I can't ever imagine doing is having AI actually write anything for me. Check my grammar---sure. I'm awful at proofreading my own stuff, as anyone who reads my blog knows. Sometimes my prose is downright unintelligible because of leaving out words that change the meaning of my sentences. For that I can use AI. But to complete thoughts? To formulate ideas? Maybe for something mercenary like a resumé.
As for physics, let's see if I have the stamina and courage to see through my ideas. Does it matter? Somehow it does to me. One good paper could make up for years of dormancy, the years the locust ate up, as the Bible says.
No comments:
Post a Comment