For a long time, I have promoted the use of talented generalists, like myself, for technology positions. However, I’ve noticed that recruiters have no way of factoring in talents. They are using old fashioned search techniques to come up with specific skills to match lists presented on requisitions. The result is a set of candidates that have specific skills, but very little breadth or flexibility. As the business changes quickly, the specific-skilled-employees don’t adapt, IMHO.
Way back in the dark past (the 70s and 80s), we were investigating the use of AI techniques to build programs that could help with automating diagnosis of computer and network problems based upon a growing database of actual solutions. After a lot of failure to write the correct heuristics, we fell back on Bayes formula to create a set of questions pulled from the database of solutions. The questions were formulated as “Did you see X symptom?”
The database was set up such that the computer could calculate that for A number of times the symptom occurred, the solution B was correct for another certain number of times. Bayes turns this into a calculation that spits out that “The probability is Z% that B is your solution.” The solutions were ranked by most to least likely, and presented as a checklist of solutions to try. It was a fun and interesting project.
Google uses Bayes and a lot of other math to drive its search engine as well as the page ranking. Match.com does much the same. There are probably others. They live with sloppy results, probabilities instead of absolutes. Why does recruitment look for absolutes?
My suggestion is their response will be: “That’s the way we’ve always done it,” or “That’s what the customer demands.” Well, with the thousands of technology jobs going unfilled from recruiting ineffectiveness, I suggest we start doing things differently, and start teaching the customer to do the same.
Your rants are gratefully accepted.