T@W Weekly: Imposter at Work

The adaptability quotient, LinkedIn unifies platforms, and WeWork positions themselves as tech

The Word: A Deep Dive Into AI Bias

In case you missed it last year, Amazon had to scrap an AI tool it was using in hiring when it showed bias against women. Reuters broke the story in October 2018:

“Everyone wanted this holy grail,” one of the people said. “They literally wanted it to be an engine where I’m going to give you 100 resumes, it will spit out the top five, and we’ll hire those.”

But by 2015, the company realized its new system was not rating candidates for software developer jobs and other technical posts in a gender-neutral way.

That is because Amazon’s computer models were trained to vet applicants by observing patterns in resumes submitted to the company over a 10-year period. Most came from men, a reflection of male dominance across the tech industry.

Whoops! While Amazon eventually disbanded the team, now almost two year ago, the lessons should still ring true today. Yet, are those lessons being heard and learned?

So yes, you’ve heard a lot about bias being integrated into AI. I may be a source for some of that POV. For people who are building and managing workforces though, it’s a really important discussion and consideration.

That’s why Julien Lauret’s teardown on the Amazon AI fiasco should be required reading for anyone looking to add AI to their hiring process. Is it a long read? Yes, 20 minutes — a lifetime on the internet. But the conclusions are sound — we need to get beyond simple technical design questions:

What does that mean in practice? In the case of hiring — rating and ranking human beings — it can be very problematic. The engineers can’t hide score calculation and weightings in the black box of the model. All the company’s goals need to be made explicit and comparable to each other. For example, let’s say that a female employee is considered twice as valuable than a male employee in a technical team. The team would need to document that parameter explicitly in the software code and its accompanying design documents. Whether or not that’s an acceptable option comes down to legal, ethical, and business factors. One thing is sure: it goes much beyond technical design choices.

This isn’t to knock Amazon uniquely. The fact that this was happening in 2014 shows they were simply making mistakes three to five years ahead of the broader HR technology market.

The real point here is that I don’t know if we’ve truly learned how to manage this inherent risk.

What the Click?

  • LinkedIn announced it is moving its core talent products (LinkedIn Jobs and Recruiter) to a single platform. A unified platform and dataset will be a big win for users and hopefully allow them to add on more capabilities.

  • David Bernstein writes in Recruiting Daily about recruiting automation and chatbot progress. As an aside, he also highlights the continuing consolidation along the fringes of AI-powered solutions in talent.

  • Plenty is written about engaging employees but how about the rest of the workforce? HR Technologist covers how contingent and contract workers can be shown the love as well.

  • The Degreed blog covers the history of corporate learning and development, though their claim that my Super Smash Bros and Third Eye Blind knowledge is irrelevant is sorely incorrect.

  • Workhuman (formerly Globoforce) announced revenues of $500M in 2018 and are forecasting $700M in 2019 — big growth in the employee recognition space. They continue to split employee growth between Dublin and Boston.

  • Last week, I mentioned that IQ was out for CQ (C is for culture). This week? AQ (A is for adaptability) is in as the BBC covers why adaptability is so important for the changing workplace.

WeWork The Tech(?) Company

Last week, WeWork (or, as they are now called, The We Company) filed its S-1 paperwork to go public and it filled the internet with some cutting commentary on what is and isn’t tech in 2019. The We Company is in the real estate business: They spend most of their money on property leases, which are then resold. Yet, their filing is filled with techno language. As Elizabeth Lopatto for The Verge writes:

WeWork — excuse me, The We Company — is primarily a landlord for freelancers and companies. You pay rent on your desk or whatever, and then you don’t have to work in the same place you live. (There are also conference rooms.) And yet the word “technology” appears 110 times in the S-1. “We provide our members with flexible access to beautiful spaces, a culture of inclusivity and the energy of an inspired community, all connected by our extensive technology infrastructure,” The We Company tells us. But I am having the damnedest time figuring out what the “extensive technology infrastructure” is. Does this just mean Wi-Fi? Is it the neon lights? Is it… lasers?

It’s true that The We Company hires lots of engineers, product designers, and so on. But, like, what major company doesn’t? If that’s the standard by which one considers a tech company, JPMorgan Chase is one of the biggest and most important tech companies on earth.

While software-as-a-service multiples in IPOs are what every company wants, how far can companies run from the very nature of the business?

The only thing that We has in common with many SaaS companies is a lack of profit. What it doesn’t have in common is the outstanding long term lease obligations: more than $47B.

As Scott Galloway writes, the high valuation driven by technology and Softbank’s generous valuation may be vaporware:

The last round $47 billion "valuation" is an illusion. SoftBank invested at this valuation with a "pref," meaning their money is the first money out, limiting the downside. The suckers, idiots, CNBC viewers, great Americans, and people trying to feel young again who buy on the first trade — or after — don’t have this downside protection. Similar to the DJIA, last-round private valuations are harmful metrics that create the illusion of prosperity. The bankers (JPM and Goldman) stand to register $122 million in fees flinging feces at retail investors visiting the unicorn zoo. Any equity analyst who endorses this stock above a $10 billion valuation is lying, stupid, or both.

If you’ve been to a WeWork space, it’s easy to get the appeal. When I was based in San Francisco, they were everywhere and they certainly beat an old Regus office space that seemed transported straight from 1992.

Still, they are ultimately in the real estate business and they have plenty of competitors. A hot desk at a local coworking spot is $175/mo versus the $310/mo you’d pay for the space at WeWork. Other than vanity metrics, it’s hard to say that the fruit water and networking is worth nearly double the price. But it certainly isn’t the technology that’s driving WeWork forward.

The We Company will ultimately push the limits for the “show viability and let someone else figure out how to make it profitable” ethos of Silicon Valley.

T@W Playlist of the Week

Here at Tech@Work headquarters, it’s feeling a lot like the end of summer. We’re getting ready for the first day of school and those long days are getting shorter and shorter. I’ve been listening to this chill playlist before we get into rain jacket weather here in the Pacific Northwest.

And Finally… Imposter Syndrome

If you’ve ever felt like you don’t deserve a job, don’t belong in a room, or are a step or two away from being discovered as a fraud, you may have something called imposter syndrome.

For me, typically before a big meeting or presentation, I have to remind myself that I’m not a fraud and have important things to share. Sometimes this pep talk even works!

What research has shown is that these feelings actually affect women and minorities disproportionately. Last week, I found The New York Times guide on how to overcome imposter syndrome incredibly helpful, even though I’m neither female nor a minority.

I’ve started talking more to colleagues about when I’m feeling uncertain about something. In reality, I’m trying to gauge whether it’s something that lives in my head or if there are legitimate issues I should be fixing. And like others in the article mention, you’re likely to swing from times of high ego to the lowest of low egos, making it even more difficult.

I’ve always joked that confidence starts with a con, the most important of which is to con yourself into self-belief — in spite of evidence of past failures.

This guide helped make me feel less like I had to trick myself, and more like I have that capability. While it wasn’t written specifically for me, I’m glad that it’s out there for anyone in the same position.

Cheers, Lance