Gergely Orosz’s Post

LEGO welcomed two new board members today. Here's me meeting with one of them this week: Ebi Atawodi! Huge congratulations!! Ebi is hands-down the best product "getting things done" person I know and have worked with. At Uber, I was lucky enough for her to be our team's product manager. We knocked it out of the water. Ebi got the whole engineering team involved in product decisions, had regular 1:1s with several engineers on the team, and we very quickly became a team of "product-minded engineers." (Here's a longer post about product-minded engineers: with a lot of inspiration coming from this exact team! https://lnkd.in/et7nWBgW) After Ebi's scope expanded (and she was no longer only working with our team): engineers on my team proposed projects to build (bringing a business plan to back their thinking) and we got those done. A memorable project was an annoyingly vague payments error message we kept seeing, but ignoring for years. An engineer estimated that allowing to fix the most common underlying errors with a tap (expired credit cards!) would bring in $1M/year in annual revenue. We made the fix, and it was a $20M/year incremental increase (!!). This project ended up being the second-most-impactful project (by revenue) the team shipped that year: and it all happened because engineers were encouraged, empowered and knew the tools on how to propose projects they thought were important, in a way that they were likely to get prioritized by the business. If you'd like to get a sense of how Ebi thinks, this episode with her in Lenny's podcast is an excellent place to start: https://lnkd.in/eDBeyTsz And see the press release from LEGO here, with the other new board member, Ilkka Paananen (cofounder and CEO of Supercell) also introduced: https://lnkd.in/ehRC7GYU. Can't wait to see the impact you'll make at LEGO, Ebi!!

Huge news for anyone working in tech in the US: noncompetes will be banned: not just in California (like before), but nationwide. This is very, very relevant for anyone at Amazon (which is the Big Tech that has enforced noncompetes even for low-level engineering positions). But it's just as relevant at other companies that (outside California) added noncompetes to contracts. Other countries should take notice. The FTC has correctly determined that noncompetes is bad for the economy: although undeniably good for businesses that want to keep wages lower, and enforce lower attrition. If you read the ruling closer: there is an exception where noncompetes can remain for executives. The regulation defines as an executive as those making more than $151K/year AND being policy makers. Many senior-and-above individual contributors will make more than this (especially in Big Tech). But they are not policymakers/execs! That's usually Director-and-above. The regulation is expected to be in effect in a bit over 4 months' time. During this time, organizations can sue the FTC to get this reversed: and the US National Chamber of Commerce has immediately announced they will do just this. Still, there's now a very real chance that soon, noncompetes will be a thing of the past for almost all US workers. We've seen what happened in states that did this earlier: California is the hotbed of innovation and startups. It also has a ban on noncompetes. Coincidence? The FTC doesn't seem to think so. Other countries (that still have noncompetes allowed) could well take notice. The FTC ruling source: https://lnkd.in/dFeVcXwr

Interesting reflection on "hypergrowth" (hiring a large number of people after raising a large funding round) by Tuomas Artman, who went though all of this at Uber: (Tuomas is the cofounder and CTO at Linear. From him:) "At Uber, I learned enough about hypergrowth in headcount that I never want to experience such a level of headcount growth again." Here is the typical path he's seen for companies that find product-market fit, while raising capital aggressively: 1. Try and find product-market fit (PMF). Try out a bunch of stuff. 2. Found PMF. Hooray! 3. Scale, fast: partially by hiring. 4. Hire *lots* of people! 5. Firefighting, because you don't have much time to invest in infrastructure (need to keep up with the competition, and also onboard) 6. You (finally) get room to breathe! Usually a year or two after you hoped you would. 7. Rewrite it all. 8. Migrate and rebuild that will take years So, back to hypergrowth: "The big problem with hypergrowth in headcount is that after a while, barely any of your engineers work on the product. I noticed that a few years into hypergrowth, it felt like 70-75% of engineers worked on infrastructure, and no one was really working on the product. Having a large number of people work on infrastructure is one of the major things we at Linear want to make sure that we do not do. We want to keep software engineering as a product problem. Every single engineer should be focused on making things better for the customer, not having to worry about the infrastructure." And hence the deliberately slow hiring pace at Linear! The above was an excerpt from the story of Linear, published in The Pragmatic Engineer. Read it here: https://lnkd.in/eqxyrh4M

I had no idea Amsterdam could have so many engineering leaders/managers come together! ELC (Engineering Leadership Community) is shaping out to be a great group here. It was a nice event. If you're an engineering leader (or close to making this transition), I recommend becoming part of this group (it's free!) and look out for upcoming events here. There are ~monthly ones!

There are truths that are widely known by insiders; not particularly a secret; and yet rarely ever written down. Here is one: Hiring processes for CTOs, Head of Engineering, VPE or Sr VPE are all bespoke/different between companies. To the point of wildly different from how the company hires for any other non-executive role! I’ve been talking with folks interviewing for CTO and Head of Engineering positions the past few months, and this is absolutely what they are seeing. The only shared characteristic is how these days processes are longer than before. Outside of that: there are processes that are 10+ interviews long (with more and more people brought in); ones heavy on references; ones where the board vetoes candidates / pushes candidates through and so on. There is no “standard” process at those levels. Here’s this articulated in Will Larson’s latest book, The Engineering Executive’s Primer: (I’m enjoying the book so far. Get it here from the publisher: https://lnkd.in/e8KabBaS )

This week, X (formerly Twitter) shipped the most amateur URL rewrite I've seen, and created a massive security attack vector. Here is what went down: Twitter has been rebranded as “X,” but the twitter .com URL remains. In an attempt to fix this, starting on Monday (8 April,) X automatically changed anything in an URL with twitter ․com to x ․com. This would have been fine, if it would have only changed twitter ․com to x ․com, almost like a permanent rewrite. However, the change replaced all twitter ․com strings inside of any URL to x ․com, while still redirecting to the underlying URL. So, for example: space-twitter ․com was displayed as space-x ․com (but directing, upon clicking, to space-twitter ․com) Sounds pretty innocent so far, right? After all, SpaceX surely owns SpaceTwitter ․com as well, so that it’s hard to hack SpaceX .com. But consider a malicious attacker wanting to hijack the Netflix ․com domain. All they now have to do is to register the domain NetfliTwitter ․com, and post a link to it! Users on Twitter will see Netflix ․com displayed, click, and the attacker can attempt to steal their Netflix passwords (in the screenshot below) X just opened up a honeypot for phishing attacks by neglecting very basic security principles. This mistake suggests a couple of things: 1. The engineer(s) making this change are inexperienced. (I cannot imagine an experienced software engineer who wouldn’t consider security vulnerabilities with URL rewrites.) This is ok! The problem is how such a major change was most certainly not reviewed by a peer, an experienced engineer, or a security engineer. 2. Either X has no security team, or a security team never reviewed this change, or the security team is useless. 3. If high-profile changes like URL rewrites ship without concern for security practices, what kind of attention to detail can we expect for less visible changes? X has closer to 500 million monthly users. You would expect basic security hygene at such a company. What is certain: this kind of miss would have been less likely to happen at “Twitter 1.0.” That version of the influential social media platform invested plenty in planning and security, meaning it’s likely this issue would have been caught and resolved before shipping. To its credit, X reversed the change in one day, and by Tuesday (9 April) it was rolled back. The fact that such a widespread change was released with no feature flags and no staged rollouts to hundreds of millions of users, suggests the culture of shipping to production at X can be described as a “YOLO strategy”. I’ve covered the various strategies in shipping to production, and more resilient alternatives to YOLO here: https://lnkd.in/dsQzhQ7S --- This was an expert from yesterday's Industry Pulse in The Pragmatic Engineer at https://lnkd.in/e2BzQCxZ. Follow me for similar updates on analysis of interesting software engineering-related events, or sign up to get it in your inbox, weekly: https://lnkd.in/grXSBkVw

The Pragmatic Engineer Job Board and Talent Collective was reasonably successful by all conventional metrics. And yet, after 2.5 years of operation, I am retiring it for good. Here's what happened: Short version: The tech jobs industry is changing, a lot. Job boards are becoming more and more useless/noisy. The "best candidates" are no longer signalling their availability. Looking ahead, platforms like LinkedIn Recruiter look to be winners, alongside specialized tech job firms. Business was ok, but on a stable decline. This move was kind of overdue on my end. Longer version: The trigger for all of this was how the vendor I was using to run the job board and talent collective - Pallet - retired the platform I used. They decided to pivot to a search firm approach. For everything I understand, this was absolutely the right decision from them. Pallet started out as a job board. However, early 2022, the startup innovated with the concept of a "talent collective." The core insight Pallet had was how people trust "creators" – and how these "creators" already have engaged audiences. I was personally reviewing every candidate application: and for the ones that I approved, I added a note on why this profile is interesting or standout – often adding details that I assumed some recruiters could miss (such as open source contributions, mentoring outside of work, or interesting and relevant extracurricular activities that I found encouraging to see.) My job board had an (initial) advantage of lots of qualified candidates applying. My job board was not a standalone board: but it was interesting tech jobs listed at the bottom of my blog and newsletter. As such, companies initially received lots of qualified candidates. Still, the majority of companies working with me never posted a single job! They went for the Talent Collective instead and contacted vetted, senior+ engineers there instead. About 80% of my customers were like this. So why did Pallet decide to shut the board down? It was probably down to two broader trends: #1: "Creators" stopped running their Pallet job boards and talent collectives. In early 2022, YouTubers and newsletter writers started to run their Pallet boards by the dozens. As I look back now, about 90% of the people I remembered having a Pallet stopped maintaining it and using it about a year ago. #2: Interest in job boards/talent collectives declined, even as traffic displaying these offerings increased. Here are statistics on traffic that ended up on my job board (when people clicked through from my newsletters and blog): and how traffic across my newsletter and blog changed over the same time: There are a few more reasons - including how automated AI application tools will, most certainly make job boards less useful, and others. Continue reading on my blog: https://lnkd.in/eZnhqRzu This is the reality of experimenting with different things: some work, until they don't. Like now!

I sometimes get asked: “do you do career coaching?” I do not: but Keki (whom I worked side-by-side at Uber) does! So inspiring to see these stories from her career development business:

When were you born? Asking to get a sense of what "generational cohorts" are most active here, on LinkedIn:

WOW. Six months after release, The Software Engineer's Guidebook has sold more than 25,000 copies. This is far beyond what I hoped for - especially since it's a self-published book! It's funny how things work out in the end. I *really* wanted this book to have a "proper" publisher and spent 6 months pitching to "traditional" tech book publishers and working with them. However, 2 our of the 3 publishers I originally pitched this exact book passed on it, and the third one pushed for changes that would have drastically altered the book in ways I disagreed with. I made the (at the time, very difficult and risky) call to break the contract with this publisher, and write the book in the original form I envisioned it - and the only way I could see it work. It's been weighing on me since: what if that publisher was right, and the concept would not work, and their well-meaning suggestions to "dumb down" the book would have been the way to go? I'm relieved that this was not the case, in the end. The audiobook is now in the works, as well as translation to 7 languages (German, Korean, Japanese, Simplified Chinese, Traditional Chinese, Hungarian and Mongolian). Thank you to everyone who has picked up a copy of the book (or the e-book!) This is all thanks to you 🙌 More on the book here: https://lnkd.in/gq2722ZF