My 32nd week studying web development at Learners Guild in Oakland, California, ended on 15 December, so I have 8 more weeks before my 40-week program comes to an end. That’s when my role changes from Learner to something else.
That something else, according to the Guild’s model, is remunerative employment. The Guild actually cares about Learner success in getting employed, because the Guild’s investors are fronting the full cost of this vocational education and get a return on that investment only in proportion to the post-graduation income the ex-Learners make. I care, too, because the Guild is trying to create tech opportunities for marginal groups (like us geezers) and I hope to do my part by making enough money to more than pay its investors back for the risks they have taken.
So, I’m beginning to glance at the job market.
One thing I notice is the state of information technology in that market. Numerous entrepreneurs are busy trying to make the market more efficient and squeeze part of that efficiency increment out as revenue for themselves.
From my perspective, so far they are doing a horrible job. Most of them are obviously using algorithms to solve the “assignment problem”, i.e. the problem of matching people on the two sides of the hiring transaction. That’s only natural, but the abysmal quality of the algorithms isn’t. In a world where experts are debating how many decades will elapse before machines become generally smarter than humans and in which machines already are smarter in some diagnostic, game-playing, and other arenas, it is shocking to see big players in the employer-employee assignment space using algorithms as crude as those I have witnessed.
Here’s an example. One important category in my industry is “front-end development”. That refers to building software that makes your web browser do the fancy things it knows how to do.
So, naturally, matching services that operate in this industry need to classify jobs as to whether they are jobs in front-end development, among many other things.
Along comes Indeed, one of the premier providers of this service. Like its competitors, it lets job seekers create profiles and category lists and then notifies them of matching jobs. Today I got such a notice, titled “23 new Front End Developer jobs in San Francisco Bay Area, CA”. Pretty clear. I expect to see a list of 23 such jobs. Here’s one of them:
Support Manager Front End Lowe's Inc. 21,819 reviews - San Francisco, CA Position Description The Support Manager role will have a focus in one of three areas: Front-End, Back-End, or Night Operations. The Support Manager Front-End role is primarily responsible for planning, scheduling, monitoring, and successfully implementing all non-selling operations in the front-end of the store (i.e., cashier and administrative functions), facilitating the store’s ability to provide a superior customer shopping experience and maximize sales and profitability. This includes overseeing the Administrative office, researching shortages or overages, depositing cash in the bank, handling register pulls and loans, managing exchange and loaner accounts, and monitoring Customer Service desk activity.
In addition, my profile clearly tells Indeed that I am qualified only for what are called “junior” software development roles, but 12 of the 22 that are actually for software-development jobs are explicitly for “senior”, “principal”, “lead”, or “director of” positions.
If this were the state of the art, one might wonder whether Indeed is rational when it uses AI in lieu of human classifiers. But it isn’t the state of the art, so my wonder is why it can get away with this level of accuracy and still compete effectively in its market.
Maybe the answer is in the quality of its competition. Look at Localwise, a small Bay Area outfit that has entered the same market, and that the mayor of Berkeley touted a few days ago. It lets the visitor choose a job category and then shows the visitor matching jobs. Try choosing “Startups & Technology”. Look at what you get: “Happiness Ambassador”, “Tour Guide”, “Delivery Driver”, “Babysitter or Nanny”, “Sales Associate”, etc.
“But maybe all these employers are startups”, you might say. Good point, and, if true, this would imply that Localwise is using a dysfunctional classification. But it isn’t true. The sales-associate job is at BioAssay Systems, which says it has been in business 14 years.
Moral of the story? We can draw our various conclusions. Mine is that there is still plenty of room for those who have common sense and a zest for building quality products to break into this market (and many others).