Mitchell Hashimoto takes on a new individual contributor role at HashiCorp | Hacker News Reader

Mitchell Hashimoto takes on a new individual contributor role at HashiCorp

hashicorp.com

357 pointstosh70 comments

As an employee of Hashicorp, I'm very excited for this change. Not that Mitchell was anything less than a great leader—on the contrary, I thought he was excellent. It's just that he's also got such a great ability to identify the right abstractions when tackling a problem. The fact that he'll be "in the trenches", so to speak, gives me even more confidence that Hashicorp will continue to make tools that strike a nice balance between power and simplicity.

How do you like working at Hashicorp ?

Frankly, I love it. The attitudes that have made the open source community welcoming are the same ones that permeate the company. Generally speaking, people are very friendly and helpful. That said, it's a job and your mileage might vary. Personally, it's the best place I've worked, and I don't plan on leaving for many years.

Not orangepenguin but agree pretty much every engineer in the company is stellar, approachable and helpful. So much to learn from so many people.

How much you enjoy day to day will depend on your team and manager.

I would return in a heartbeat to work with the folks building HashiCorp cloud.

Well done. A very graceful and transparent announcement.

I decided to leave senior leadership and return to IC engineering work last year. I was met with shock and, frankly, negative feedback from my peers. It has been a difficult transition, and it's easy to miss the benefits of leadership, but I knew engineering was the right place for me.

It was difficult to do this as a veritable nobody, I can't even imagine the difficulty of the decision for Mitchell. I wish him luck and continued success.

People really hate the idea that leadership isn’t always the best place to be. “But with leadership you multiply your impact!” Yeah yeah yeah.

I just like building things, convincing people of what should be built is another matter entirely. There is so much stupid horseshit you have to cut through, due to the multiplicity of interests in play at any given time. I’d rather fight with a buggy framework or poorly designed toolchain or whatever.

Well said. If you look into some of the research and theories on leadership, you find a very different picture. You could someone a leader any time they have influence over people toward a goal. With this in mind, you can see that an "individual contributor" can definitely be a leader.

However, in business, leadership is often conflated with "legitimate power" (as opposed to expert or referent power), meaning that the person holds a title and is assigned direct reports. Titles and employees are not necessary for leadership.

I wish that engineering departments were better at highlighting this. A person with practice and experience can be a leader by simply working hard, setting a good example, giving input in team meetings, etc.

Such a great guy. At Hashiconf a few years ago he was just one of the normal people, a load of us all went to the speakers’ dinner and he just found a seat and sat down - opposite me, as it happened, and we had a laugh. Armon is a top chap too, we would talk motorbikes if I ever bump into him. Their tech skills are top-notch and I wish them the best of luck. So many tech companies would be better off this way.

It's strange, it feels this way at Hashicorp, too. I was a little shocked when I was new to see Mitchell occasionally commenting on RFCs, messaging people to ask questions, and updating documentation. If you didn't know his title, you would have thought he was just another engineer. He doesn't elevate himself above others, and Armon is the same.

I honestly think that sort of kindness and respect is actually what allowed Hashicorp to thrive in the open source communities.

Yeah, Armon is a sweetheart too and both of them have put the legwork in to meet and understand the needs of folks like me, and everytime there's a new product I get excited to use it. Their code has literally given me a livelihood for six years and made hard things easy. I've met loads of Hashicorp folks over the years and they're a top bunch.

Having Mitchell comment that my RFC was good was a huge highlight of my tenure there. He deserves all of the success he's earned.

Same here. I was sitting next to them at an after conference event and had a good laugh. Only later I recognized their faces. Or having a beer at FOSDEM.

I've been a user of their products for years, and really like them but never knew the last name Hashimoto seems to be why the name of the company is what it is. <Mind Blown>

It is! Most people pronounce the first part of Hashicorp the way you say "hash" (as in MD5). Technically it should be "haw-shee" because of the last name Hashimoto.

Side note, internally a pseudo company name used for example purposes is "Dadgarcorp" after Armon Dadgar. :D

Stepping down from a leadership position in order to focus on engineering is something as rare as a unicorn. This person has my respect, as always.

zero ego, just trying to find the best position to make the largest impact, love the move and the mark of a true engineer.

Mitchell was a prolific coder even when he was the CEO. The public Hashicorp repositories have his fingerprints all over them. For the recently launched Waypoint, which is delightful product, Mitchell was the lead contributor, by far.

Mitchell was a prolific coder even when he was the CEO. The public HashiCorp repositories have his fingerprints all over them. For the recently launched Waypoint, which is delightful product, Mitchell was the lead contributor, by far [0].

[0] https://github.com/hashicorp/waypoint/graphs/contributors?fr...

What Got You Here Won't Get You There is a frequently recommended book about changing yourself to get to wherever There is. It seems that Mitchell has chosen to stay the same and change where There is instead. Sounds like a fantastic way to live.

It is sort of strange to read about it.

>I founded HashiCorp in 2012 and served as CEO until 2016,

>I'm incredibly proud that as an executive, I helped HashiCorp grow from nothing to nearly 1,500 employees with a valuation of over $5 billion.

It wasn't that long ago Vagrant was "the" tool for the job. Now it is a $5B valuation company! I hope HasiCorp can continue to grow and wish them all the best.

I still have my Hobo sticker somewhere...I mean "vagrant" :/

This is one of very few tech S1's that I am looking forward to reading and investing in. They write beautiful tools and beautiful documentation and I believe in their team.

Reassuring to know that there is a path to transition back into an individual contributor role.

If anyone has lessons learned or thoughts in that direction I'd love to learn more.

I think there always is if you can accept it yourself.

Sometimes. When a previous manager quit, he told me he had been a manager for a year and then tried to stop being a manager for another year. No path back to IC at our company so he joined FB as a staff eng.

I’ve seen this happen a lot in larger companies where people that got promoted to managerial posts revert back to ICs after a while. But definitely not at this magnitude; stepping down from executive team is probably unheard of.

I went through this. It was one of the hardest things I've done in my career, but it has proved to be the best strategic move I have made. I am much happier and my scope of influence in our organization is still wide, just more technical.

It really depends on your team. I hired my own boss to replace me, and worked with him on transitioning me out of management. As the new guy, it was in his interest to keep me happy. Worked out really well.

Totally depends on the the people though - your boss, your reports, the senior leadership team.

This is a really interesting move, and certainly one I've never heard of. As a founder/CTO it's something I'd love to do one day. Are there any other examples of this occurring?

Orion @ Heroku. He even left shortly after the acquisition, and then came back as an IC have an extended sabbatical. Possibly Lee Holloway @ Cloudflare while he was still there? I don't know much of the internal workings there to know if he was ever considered part of management or was simply a co-founder IC while he was there.

There are 2 aspects to this. 1. The importance of having a solid technical contributor in an executive role cannot be overemphasized. 2. If you were not a CEO/CTO before it's easy to get anxious about the decisions executive leadership is taking and not having a say in it. I mean it might work out just fine for Mitchell because he was the CEO, so he might still be involved in major decisions or at least his opinion will be actively sought. Or may be you just enjoy doing what has been told to you and that's totally fine too; as long as it's a conscious, well thought decision. But if you've any opinions on how things should go, it's going to be a tough sell in a non-exec IC role. My point being this is a specific case and it'll work out great for Mitchell, but for others think from all angles before making this kind of move. Obviously, if you really hate your exec job or are not very good at it, you're better off moving away from it.

I really dislike the term "individual contributor".

I know it's an industry standard term for anyone who doesn't have direct reports, but it feels very inaccurate to me.

Just because an engineer doesn't have direct reports doesn't mean they aren't exhibiting all kinds of leadership behavior that elevates their contributions above the level of "individual" - teaching, mentoring, strategy work, leading projects.

I don't have a good suggestion for an alternative term though.

Think of “contributor” as the key word in the term. It covers all that leadership.

Then note “contributor”’s absence from “manager.” :)

Maybe technical contributors? You're right. In the most efficient organizations, leadership lets the technical decisions bubble up from the experts, rather than dictating them from the top.

> Just because an engineer doesn't have direct reports doesn't mean they aren't exhibiting all kinds of leadership behavior that elevates their contributions above the level of "individual" - teaching, mentoring, strategy work, leading projects.

That's a good point, but in my experience "individual contributor" is just a code phrase for "not managing direct reports".

Being an IC doesn't mean you're not contributing leadership or strategy, it just means that you're not doing hiring/firing or performance reviews of direct reports.

Yeah that's how I understand the term too. So it's an effective piece of language, I just dislike it!

Yes, but at a lot of larger companies and perhaps the vast majority IC is code for "them".

"Us" is management BTW.

As someone who has also been on the journey from early employee into project and team leadership, a formal management role, and then back to individual contributorship, I kind of agree, though at the same time I also don't really care about the title, and I know that being a "senior" or "staff" level individual contributor clearly implies a bunch of those other pieces, including things like soft leadership, independence to pursue strategic initiatives, etc.

Yeah, titles like "staff engineer", "principal engineer", etc. are good because they imply expertise rather than responsibility for direct reports.

Eh, it's kinda like the term "stakeholder." It just reeks of bureaucratic bullshit, but otoh it's a very real important concept that we just don't have another good word for, so we stick with it.

I agree about the IC terminology.

The best I've seen so far is "maker" vs "manager"

Ooh, I think "maker" would be a good word for it, though it's pretty inextricably tied to hobbyists.

I've definitely been in a company where they referred to me pointedly as an IC, as a way to signal they were not interested in me taking on tech leadership roles. I got the fuck out.

Has this ever happened in another software company? I can think of Sergei Brin, but did he really produce after exiting CEO? Or David Filo of Yahoo with the same comment.

John Carmack did something kind of similar a billion years ago at Id.

John Carmack and Dave Cutler have always been individual code contributors, regardless of their titles and management responsibilities.

They never cut out management/executive duties completely (as in Mitchell's case), but they came somewhat close.

Carmack has been pretty explicit that he is not, nor does he want to be, a manager (either in the executive sense, or the “technical lead” sense). He sees himself more as a senior craftsman.

I hope he is creating a simpler Kubernetes competitor for robust container orchestration/management. Hint hint.

Do you mean Nomad?

I did not know about this: https://www.nomadproject.io/docs/nomad-vs-kubernetes

Is it good? Is there a growing user base? Why do I not hear about it?

I really like it. K8s seems to have won so I'm always worried Nomad will be shuttered but I'll use it as long as it's supported.

I am fully in bed with Kubernetes professionally and do love it, complexities and all.

However, I will say I feel the Nomad stack is quite good and fills a nice space where full-blown K8s is way overkill or just not the right tool.

It fills another weird niche which is where Kubernetes can't scale _enough,_ especially in large hybrid environments.

We have several thousand hosts in a single cluster across Windows/Linux, all hooked up to Nomad/Consul/Vault and it takes some effort but our team isn't that big and we manage it. With the most recent releases of Nomad and Consul they truly shine at both ends of the spectrum, small/simple and extremely large/extremely complex.

Its working pretty well for our 8 person team. We started with ECS, but at some point it wasnt enough, and kubernetes would be too much for our size and expertise (We have around 100 services and do not need autoscaling).

Nomad/Consul/Vault/Fabio are super easy to setup, and its like 10% of the effort of handling your own kubernetes.

Documentation could be worked on, but its getting better.

we canned our Kubernetes implementation after actual years of trying to glue together tools at a medium sized company with a small engineering team. we switched to Nomad and are very happy.

I am of the belief that there's an inverse bell curve of support for Kubernetes coming from 1-3 person teams and large companies. either you've got a 1-3 person engineering team who understands the whole thing, or an enterprise that can support it with a team of SREs helping keep the manifests in order. anywhere in between can become chaos quickly if you can't support training your department on it.

I honestly don't know why more people don't hear about it and use it. Probably just hasn't hit that "critical mass" yet. I haven't used all of the tools out there, but I've used ECS, K8S and Nomad. Nomad seems to provide the most "efficient use of complexity". What I mean is, for a given increase in complexity, you get more utility out of Nomad than out of the other tools. You don't feel like you're fighting it, or forcing it to do something it wasn't designed for. Even in disaster recover situations, Nomad is surprisingly simple to manage.

Is it to early for a technical wishlist? I will just try :)

A HashiCorp solution for something like renovatebot/dependabot but with the special HashiCorp touch (and maybe Packer/Vagrant support) and suddenly your OCI images run on Nomad

What's wrong with Renovate / Dependabot, if I may ask?

Not that I don't like the Hashicorp touch on things, quite the opposite, but I'm genuinely interested in your thoughts here since I cannot find huge problems in those two tools (:

There is nothing wrong with them. The HashiCorp touch maybe envisions something I did not think about before.

If I ever become a CEO of a software company, I'm pretty sure I'll miss coding in months time. This is a really good example for the future (if it works) that you can still be a SE and own the company at the same time

It's interesting to see that this decision was made years ago, because it feels like he's had a much larger impact on the development of Boundary and Waypoint, to their great benefit.

Mitchell deserves all the praise and the success he’s had so far. I just want to point out that he is a “gifted” individual so if you’re the kind of person that seeks to emulate his behavior I would advise that you try to understand him more.

Can you expand on what you mean by a "gifted" individual?

“Genius” might describe it better. He wrote a ton of some of the original hashitools himself.

What a great choice!

Not go be cynical, but it sounds like he's hiding something, especially with the part mentioning he doesn't have access.

Yes, in most cases I’d suspect some foul play. In this particular case though the cynicism is probably unwarranted.

He is fabulously wealthy now. He doesn't have to work. What is the percentage of his ownership in this $5B company? 10%? More? He's achieved his goals likely. Probably will want to slowly cash out and diversity.

Do other things...

> slowly cash out and diversity

I'm pretty sure that if that were the main reason, it could have already been achieved at one of the previous rounds.

I'm an employee, and I doubt he's hiding anything. Even while working as CTO, he has kept an ear to the ground in engineering and periodically participated in design discussions and other things. I really think he just genuinely enjoys engineering work. Internally I think many of us have great confidence in our current leadership, so it makes sense that Mitchell would feel like he can hand the reigns over.

I mean, many great open source projects are still maintained by the original authors. Why would it be suspicious for Mitchell to move into a better position for doing the same kind of work?

Hopefully he'll have some time to learn YAML now /s

Reply on news.ycombinator.com