Monday, March 7, 2011

So chemists are safe from the intelligent computer menace, right? Right?

Chart authors: Autor, Levy and Murname
Credit: Paul Krugman, New York Times
While super-smart computers have been beating Jeopardy champions recently, they've been putting lawyers out of work as well. This fascinating New York Times article points to computer programs that can perform "document discovery" services that used to require auditoriums (literally) of lawyers poring over documents. More neat (and or freaky) are computer programs designed to analyze e-mail for "off-the-record" conversations, i.e. the moment when you decide to stop an e-mail chain and walk down the hall and have a private conversation.

In blogposts on this issue, Paul Krugman points to a paper by economists Autor et al. that summarizes the tasks that computers can either replace or complement humans. If you see the top right section ("non routine", "analytic and interactive tasks"), it appears that chemists ("forming/testing hypotheses") are going to be safe for now from productivity gains from working with computers. It is interesting how Krugman suggests that "medical diagnoses" may ultimately move from the right to the left column. Is Watson going to put general practitioners out of work? We shall see....

11 comments:

  1. I think medical diagnosis could really be helped by computer technology. I've known people with hard-to-diagnose problems - like Lupus and Lyme Disease - that took years to get diagnosed. In fact, they ended up self-diagnosing through places like WebMD. Once they see the right people they could get diagnosis and treatment. The standard GPs couldn't diagnose easily.

    ReplyDelete
  2. Chemistry has already gone through some of this transition. When I was a grad student, we were still searching for research papers through index volumes and paper records. There was some basic computer searching available, but smaller universities (like mine) didn't necessarily have access. Doing lit searches took hours - sometimes days or weeks. Programs like SciFinder and Beilstein are huge time savers and they probably did eliminate jobs. When I first started in industry my company had a library building full of chemists whose job it was to do literature searches for chemists in the lab. This was in the mid-90's and it was already becoming obsolete. The number of employees dwindled until the library was shut down completely a few years later.

    ReplyDelete
  3. I am old/young enough (35) to remember actually using bound copies of Chemical Abstracts in the department library during my BSc/PhD.

    The idea seems so quaint now, but back then (1997) the web was still in its infancy, Netscape was still the bomb in terms of browsers, and you actually had a toolbar button for "images" to avoid having to download such content to speed things up.

    The nature of so many things has changed so completely that the actuality is that it would be hard to imagine any area of business that hasn't been transformed over the past 15-or-so years by technology.

    ReplyDelete
  4. Kay, yes, that's very true. I remember my little company that I worked at right out of college (late 90s) had a librarian (and she had an intern for scanning journal articles into PDF (?) format...) That position was ultimately eliminated and she moved elsewhere...

    ReplyDelete
  5. Back in the late '90s, I, too, was a college student, and recall (somewhat fondly) loading money onto copy-cards and, failing that, bringing nickels and dimes to copy stores. When you wanted a journal article, that was the only way to copy it unless you wanted to take notes...most journal archives had not yet been scanned onto the web.

    Moreover, say your library didn't have that journal, and the one up the road did. In that case, we had to *gasp* walk over there, find it in the stacks, and beg those librarians to photocopy it.

    I can still picture the purple mimeograph ink of my elementary-school tests....ahh...

    ReplyDelete
  6. SAO: I had a former colleague (who I love dearly) who solved the copycard problem by purchasing his own scanner. When he went to the library, he'd bring his laptop and portable scanner and get himself a permanent PDF copy. Whattaguy.

    ReplyDelete
  7. Beilstein Crossfire came out when I was a grad student. I remember how big a change that was. Yes, one field that has dramatically changed with technology is library science.

    ReplyDelete
  8. I think there will be fewer chemists in the future and more automation. I use to work for an environmental lab, and within this company, there were stories of a lab in Chicago. One room, full of 15-20 mass specs, all run by one guy skilled at programing macros.

    ReplyDelete
  9. May be in 5-15 years' time the microbes will begin to replace chemists. You want a fancy molecule made? Just customize a microbe genome to synthesize it for you.

    Perhaps those synthetic genomes themselves can be "printed" by automated machines. The domestication of bacteria, that's going to be the end of the synthetic chemist.

    ReplyDelete
  10. @DP: An interesting proposition, but....I don't see it long-term. Enzymes / microbes can do a lot of things, I'll grant you, but most of them require pH 6.0-8.0, buffer, water, proper concentration, feedstocks, temperature control, culturing, and for the genomes to be thought up, printed out, and PCR'd just to get the whole process moving.

    Now, in case I've misunderstood the Autor / Krugman article above, all of those things I just mentioned would have to be done in a "manual labor" fashion, if for no other reason than to have someone refill the bottles and troubleshoot! So, we'll at least need a few biochemists hanging around, and (I guess) some synthetic folk. Plus, who analyzes the products out the other side for consistency and quality? AHA! - a need for QA and analytical folk.

    Add to this discussion the fact that, although we are constantly revealing new enzymes from bacteria / fungi / archaea that survive the cold of deep space, that live in heat vents, etc, bacteria are still (for practical purposes) good only at doing what they've evolved to do: methylations, acetylations, aldol, some oxidation and halogenation, cyclization, dehydration, etc. So, where's my enzyme to install sulfones? Promote aryl-aryl coupling? Fluorinate?

    People have been worried about what you've mentioned since fermentation was discovered and applied to large-scale synthesis of ethanol...about 5,000 years ago. Not to mention all those antibiotics we produce every year in huge steel vats. But (luckily) we chemists still manage to have jobs...even if some day we'll only be hauling raws in for bacteria!

    ReplyDelete
  11. Having machines / computers doing the work isn't so much a problem if you can find a way of mitigating the effects on the displaced humans. I'm sure we'd all love to have more leisure time if we didn't have the money to worry about...
    On the other hand, as all this automation requires energy, and mainly finite fossil energy at the moment, perhaps things will start to go into reverse and manual labour will be back in vogue.

    ReplyDelete

looks like Blogger doesn't work with anonymous comments from Chrome browsers at the moment - works in Microsoft Edge, or from Chrome with a Blogger account - sorry! CJ 3/21/20