markshuttle
The team working on Rust coreutils have been loving the feedback wider testing provides, and I am sure that 25.10's wide usage will enable Ubuntu 26.04 LTS to be confident in the new stack. It will also enable other distros to consider making the same change.
High school is a sort of universal misery - pretty much everyone who might be applying for a job with us, from almost any country, would have gone to high school. Rich or poor, you had to suffer through it. Yes, you're a different person now than you were then, but it's still interesting to hear how people handled work and social dynamics. There is plenty of good science which correlates young adult behaviours with lifelong outcomes. And as one part of an interview process, it's a useful reference point that is less susceptible to circumstance than things like "which university did you go to".
We now hire much from many more countries than we used to, and it feels good to me that we're giving opportunities to work on open source to a wider audience. Sure, you can be cynical about our intent. Invent elaborate motivations for our process. Perhaps the answer is as simple as this - we want to work with people who are conscientious and care about getting open source into more hands, in an easier to use form, at the lowest cost. Now, that's not the worlds most profitable software strategy, but it feels good to me to make that the focus of a days work. Terrible, right? Crazy, right?
Of course, things may change at Canonical if I am no longer involved. That's a reasonable risk to think about and have a plan for. Some paranoia is constructive. One of the nice things about open source is that you can fork it if you want to. But to do so just because Canonical might in future take a different view than we have to date seems like its paying too high a price for that paranoia :)
We've always tried to be at the forefront of new kernel capabilities - especially security and container tech - and it helps that Ubuntu generally has very modern kernels. On Ubuntu we can make releases of the kernel and LXD that line up nicely. Other distros with older kernels have always been supported as well as possible, and I don't see why that would not continue. There is certainly no plan at Canonical to inhibit that.
We've all decided to build with sustainable local materials, which anybody can dig up without harming the environment. That's wonderful! Unfortunately, the materials sometimes turn out to have asbestos in them.
Canonical and a large number of friends enjoy digging, and they make those local materials available for free, no digging required, as Ubuntu. So Ubuntu happens to be a free source of pre-dug materials which is popular for lots of reasons.
People with big buildings who like using Ubuntu need the asbestos removed if its found, and Canonical has started to do that for them commercially. Canonical have also said they are happy to remove the asbestos for free for small buildings, as long as the big-building people keep funding them to do it. The more big building people choose Canonical to sort out this issue, the better it will be for everyone using Ubuntu, even if they don't do any digging themselves.
Seems like a great deal for people with smaller buildings, as long as people with big buildings also think it's a good deal. Since I happy to like helping people with smaller buildings, I think this is all pretty square.
Over the years, companies have started asking us to do more for them in universe, and now that body of work is available to all customers. We are making it freely available to you and others under a personal subscription. I think that's rather elegant, I hope more and more companies see Ubuntu Pro as a very cost-effective way to get full compliance for their estates, and I hope we can keep growing the set of things we make available for free as a result.
If you look at the range of packages covered, and the numbers of issues addressed, it's way, way more than any other enterprise Linux offering. If it were possible to provide enterprises with this level of security update coverage for free, then I'm sure someone would have figured out how to do that. I couldn't figure out how to fund full security coverage of universe without having customers for that work. In the end, I think the Ubuntu Pro free personal subscription is a very nice way to balance what are ultimately conflicting desires between people who quite understandably want more and more for free, and people who are able to buy the work that they need.
The Canonical security team and wider Ubuntu contributors will still make best efforts to update universe. Neither the team nor I are interested in degrading a prior experience for our free users. We know what the criteria were for those best-effort universe updates, and they remain unchanged.
What's changed is that there is now a much larger team that will systematically fix every high and critical vulnerability in universe, with an SLA. That's a huge improvement, it's great for enterprise users, it enables people to use Ubuntu in regulated and mission-critical environments. It also makes me very happy that we give it free for personal use on 5 machines.
Broadly speaking, when you lead something, the people who will step up to lead alongside you will be people who mostly agree with the position you take. Obviously. Why would they align and invest precious years in something they totally disagree with? It would be dysfunctional to have a leadership team in any organisation made up of people who have hugely different opinions about everything at a fundamental level.
Of course, that group has disagreements on stuff. We spend a lot of time talking about the things we disagree on, but we live with people we broadly agree with.
So from the outside it may look like the leadership team of any company are highly aligned. If you have been at Canonical, you know that in the leadership team we have lots of hard debates and discussions, all the time, in search of the best. You also know we are all working to the same goal, which is to help people consume open source cheaply and safely, at scale, across all kinds of compute.
So, here's how I interpret your comment. You saw a group of people that were broadly supportive of a position I hold, and which you disagree with. Sure, you disagree, and that's fine. Calling them sycophants is just finding a nasty word for people who happen to be aligned on something you disagree with.
How do you deal with adversity? How do you deal with peer pressure? How do you motivate yourself to work on things that feel irrelevant? How do you choose your hobbies and the people you associate with? Those are important things to assess in a potential colleague. And this is just one potential angle on how one might assess them.
As for "making stuff up that sounds good", you are correct, some people are better at that than others. Fortunately we have clear and objective measures of that to work with so forewarned is forearmed.
I've been around for those years :)
When we need something difficult and fundamental thought through very deeply, because it has to be precisely correct, I've learned to trust our CTO. Others usually want to go faster, or fix-it-later, or we-don't-need-a-spec-just-be-agile. That creates tension. But when I think it needs to be done very carefully and correctly, some tension is worthwhile.
I wish we had been more careful with some of the foundations for Unity, for example, because they would then have served others better even if the project itself failed.
At the scale of hundreds of engineers working on tens of codebases, who hasn't?
That's my dispute with your position - it's flatly generalising and characterising ALL our work as if it was our least successful work. By all means, call us out on our mistakes. But if you think everything we have ever done was hopeless, then you must have some theory about why so much of what we do has in fact been widely adopted. Surely you can see how fragile your own arguments must be if they are at odds with some clear data.
I like challenges and criticism; it helps us get better. I know we have made mistakes, that helps us get better too. I respect people who have differences of opinion on what needs doing, or how it needs to be done. Why not take the opportunity to become a counterpart in a rich discussion, rather than a one-dimensional anti-everything-Canonical mouthpiece?
We killed Unity for three reasons. First, I ran out of money to fund it. That was rough. Second, we built some very shoddy pieces in that codebase (though we also had some teams that did amazing work). That was sad and taught me a hard lesson. Third, we were the fourth horse in a two horse race, and the telco and ISV ecosystems were clearly not going to engage.
I agree with you - the vision of Unity was and still is beautiful. It's probably right - convergence between phones and tablets and laptops is the future for laptops. I think it's fantastic that a small but feisty community continues to work with that code. I was therefor, like you, completely disappointed to fail.
But why hate on me, and Canonical, when we literally put everything we had on the line for that vision, but screwed up? Why be an ass to the sorts of people who are willing to put every ounce of strength they have into something that you really wanted to see succeed? We failed because we made mistakes, yes, but for heavens sake recognise an ally when you see one, even if they make mistakes.
A hiring lead is expected to find the best candidates, not just the ones with a higher number on a simplistic test. Your point is correct that a hiring lead with bad judgement or low engagement might get a bad result, but what hiring process cannot be undermined in that way?
When you don't like something, it's common to mischaracterize it with anecdotes of stupidity, then say it must be stupid. The argument is basically "a biased and obnoxious hiring manager could misuse that data to make biased and stupid hiring decisions". Yeeeeeees. Obviously.
The psychometric test we use is a series of numbers covering different aspects; it's simplistic, but not as simplistic as I reflected with a single number.
Hiring leads often progress a candidate that didn't ace that test because they see evidence of excellence in other data. And I often reject offers to candidates who aced that test because I think the result doesn't gel with reality, although I will usually interview a candidate if I have concerns about an offer.
When you are going to ask your colleagues to tackle a hard problem, and do it in a different way than they did before, it helps if you have actually done it yourself. If that is going to be lots of work, and if it's going to make them vulnerable to trolling on social media, even more so ;)
We started to plan growth a few years back, but seemed to be struggling to get people in. Hiring managers would say "there were no good candidates", which didn't seem right to me. Back then, every hiring manager did things independently, as they saw best. We had very little standard process other than HR onboarding.
When an opportunity came up for me to hire, I adopted a new process, then expanded that to cover a number of roles. Once the process was relatively settled, we picked a small number of senior people, and they ran this process across all roles, with a weekly meeting for us to figure out how to evolve it further.
As you can imagine, this change caused all sorts of angst. Reddit is not the only group that has thoughts :) Many of the same assumptions and opinions expressed here were expressed internally, some resigned. Systemic change in an organisation is painful, this one was particularly hard.
Now, however, most see the benefits.
Our flow of applicants has increased 2-3x. We now have 10-20,000 applicants per month, much more widely representing the world's talent. We are making more appointments than ever in our history. Our weekly meeting is now mostly a celebration of new hires - the whole hiring lead team looks at every new hire's resume and talks over the process of hiring that candidate, then discusses improvements to the process. We are able to analyse the data in our hiring process and look for any indications of bias at each stage, then engage and address it if we have concerns.
You ask if the "CEO is pushing for a candidate" biases the process. You misunderstand my role as hiring lead. I am not pushing for a candidate, I am simply running a few roles as a hiring lead. My job as a hiring lead is to do the initial lightweight screen, then let the process do the work. I don't meet candidates in my roles until the very end of the process, and then only sometimes. I'm running the process to hire for other managers, not for myself. Those managers meet the candidates in late-stage interviews. My job as hiring lead is to make sure we have a good bench of candidates before we make an offer.
Some of the roles I run, which attract the bulk of the applicants, are open-ended roles, for "engineering manager" or "engineering lead". When candidates get to me in those roles, if I think they are great I then send them to a manager who is looking for someone like that. So it has been my great pleasure to give tens of amazingly great candidates to my colleagues for consideration for roles they hadn't even started to hire for. Think how nice that feels, for me and for that manager :)
To circle back to the beginning - beware the trap of "normal". You will never achieve something exceptional if you limit yourself to normal.
Also, our process doesn't work as well for some roles and some locations, so we continue to evolve it.
But on balance, I am very happy with the growth of our teams and the calibre of our new colleagues.