Luca Rossi κ©
@lucaronin
Followers
9K
Following
4K
Media
562
Statuses
5K
π¬ Author of https://t.co/mtplxYUOk4 β’ I write weekly about making software and working with humans, to 120K+ engineers. ποΈ Sim racer @SimRacingLG
Rome, Italy
Joined November 2010
Refactoring just hit 100,000 subscribers π₯³π. This is an unreal milestone β I am incredibly grateful to all our readers for trusting us with their time and attention π we will always do our very best to deserve them. I have been a bit silent on social media in the past few
1
1
4
My TL;DR of database choices in 2024. - Pick Sqlite iff you donβt need the db to be accessed over network.- Pick Postgres otherwise.- Never pick NoSQL (and I did my PhD on NoSQL). Also shout out to this awesome article for more depth π.
74
123
1K
Here is my full chat with David Heinemeier Hansson (@DHH), creator of Rails, CTO of Basecamp, NYT bestseller author, and professional racing driver ποΈ. We discussed David's opinions on the current state of software, including excessive complexity, the role of AI, and the future
8
39
342
Some of the best-performing engineering teams DON'T use a Staging environment to test things before release. They just push to production. 1) Why?.2) How do they pull this off?.3) Should you use Staging yourself?. I talked with tens of teams and are the lessons I learned π.
7
28
196
π₯π₯ Announcement! π₯π₯. Today is a very big day for me and Refactoring, because:. β I am leaving my job to work full-time on it π.β I am adding a paid subscription plan π. A small thread about what's coming next and how we got here π
11
12
142
Here is a great list by @lara_hogan that shows how you can turn "regular", sometimes judgy questions into open, psychologically safe ones π
1
18
91
ππΌπ π±πΌπ²π π©π²πΏπ°π²πΉ ππΌπΏπΈ? π§. This week discussed this on the podcast with Malte Ubl (@cramforce), CTO at @vercel !. I loved our chat β here is what we talked about:. 1) ποΈ π©π²πΏπ°π²πΉ'π π²π»π΄πΆπ»π²π²πΏπΆπ»π΄ π°ππΉπππΏπ² (00:41)β how the entire org uses Vercel
3
14
94
Good dev onboarding in one sentence: make new hires ship production code in a week.
8
9
77
Refactoring just got to 1000 paid subscribers! π₯³. I am now officially living off my 1000 true fans!. I am extremely grateful and humbed by this! πββοΈ Thank you all for following this journey!
9
3
69
The year is closing so, as I always do, I am reviewing everything we have done at Refactoring! β¨. It's been an incredible year. We have already celebrated the 60,000 subscribers, so let me spend a few words on the community. This year we went from 200 to 600 community members,
2
0
5
400 paid subscribers! π€―. 5 months since the launch of the paid plan, @refactoringclub has crossed 400 subscribers and $50K ARR! π. Going for 1000 this year! So grateful about this journey, thank you all for your support π
10
3
62
Top performing teams I have met all keep a high degree of *alignment* and *autonomy*. This is made possible by three core elements:. π― Goals β know what you need to achieve.π Skills β have the chops to achieve it.β Principles β know how to make decisions
3
10
63
A modest proposal: letβs stop talking of 10x engineers and letβs start talking of 10x teams. Great teams create great engineers, while the opposite is not always true.
3
12
61
What is the most useful type of tests? π. One of the most popular testing models out there is the testing pyramid, made popular by Martin Fowler. The larger the layer, the more tests and effort on that layer.
3
11
61
I have struggled to draw good flowcharts for years and I just found out that there is a better way called DRAKON.
3
7
60
@paulg I am sorry Paul, but yours is some version of ad hominem fallacy. With this kind of reasoning we shouldnβt be allowed to express criticism/doubt to anyone who has achieved more than we did? . Letβs stick with arguing about current facts, not past performance.
2
0
51
When delivering features is valued more than the actual outcome for the user you end up in a bad place. @martinfowler called this Feature Devotion. Invest in a plan that is detailed just enough to support your decision making, not more.
2
12
56
Mentorship is about good answers; coaching is about good questions. It took me a lot to figure this out.
1
15
54
@GergelyOrosz I believe after you get into senior/leadership jobs the ROI of switching goes down. Leadership needs more time to create value and build trust β and whenever you switch you lose all the trust. A good pattern might be: switch often early in your career to learn more, and then.
0
2
51
The Agile manifesto is 20 years old now, but it is still very relevant. I believe there are four major themes we can take from it:. π₯ Work closely with stakeholders.π Work in small batches.π
Give teams agency.π¨ Promote simplicity and good design. I think about them often.
0
12
42
Stripe deploys 10 times a day, while maintaining 99.996% availability. This surprises many people, but it's just natural. Continuous delivery brings stability. β Frequent releases = small batches = less risk.β Fast releases = fast recovery from failure.
Fun fact: we deployed 3,350 new versions of the Stripe API last year.
0
17
48
Love it. Let me caption this π
2
5
43
The @NotionHQ AI assistant just dropped and it is *instantly* better than anything that I could ever build as a custom Refactoring GPT. It's absurd how good it is. It reliably surfaces everything I have ever written + quotes the sources, all in <2 seconds.
3
5
47
In my experience, people in tech constantly underestimate how fast smart people can grow. If you hire brilliant folks, train them and keep them motivated, you can drastically lower the bar on how much experience you need to hire for.
0
7
40
Some of the best managers I know were "generalists" that had seen many different situations. This helps them with two skills:. 1) β€οΈ Empathy β they understand what their reports are going through. 2) π° Systems thinking β they think strategically and retain end-to-end vision.
3
9
46
@GergelyOrosz Most of the companies I know that have few managers or none at all pull it off by creating distributed / automated processes around common manager duties:. hiring -> distributed.perf reviews -> rules + commissions.growth -> coaches.planning -> clear goals + autonomy.
1
1
43
Generalists have several strengths, and they all come down from a major one: they are *open*. If you look for an edge you may have over specialists, it isn't in your skills, but more in the way you think.
3
3
41
Exactly 1 year after the launch of the paid plan, @refactoringclub has surpassed $100K ARR π€― π. It's an insane milestone β I thank you all for your support from the bottom of my heart! π
6
1
42
The Agile manifesto is 20 years old but it is still very relevant. There are four major themes we can take from it:. π₯ Work closely with stakeholders.π Work in small batches.π
Give teams agency.π¨ Promote simplicity and good design. I think at them often.
2
11
41
30,000 subscribers! π₯³. Two years in, it still feels surreal that I am able to do this for a living. Thank you all for following on this journey and trusting me with your time and attention π
11
1
41
@levelsio I get your point but honestly i am not sure I would trust a democratic vote on ETF holders more than the board of vanguard etc. I mean people buy ETFs to delegate stock pick, they donβt even know what individual companies are in β does it makes sense for them to vote?.
0
0
40
I have learned this the hard way, more than a few times. I believe the fallacy is: if you enjoy doing something 1 hour/day, it doesn't meanβat allβthat you will enjoy doing it 8+ hours/day.
1
5
37
@david_perell There is a similar building in Rome that doesnβt get much credit, but I think itβs amazing. Itβs the BNP Paribas HQ
1
1
35
π₯ Introducing the Refactoring Podcast! π₯. Big newsss!! I am launching a brand new podcast for Refactoring!. Every week I will interview a world-class engineering leader about making software and working well together. Here is a peek at the first season π
8
2
37
There is a common belief that top talent requires little management effort. So if you keep hiring A-players, you wonβt need managers, or only very few. In my experience, this is totally false.
1
6
38
As a manager, you should cut meetings whose value is lower than the opportunity cost of attendees' time. This is tricky to judge, though, because ROI of ceremonies is usually not linear. Most often, a few instances repay the cost of the whole process.
1
9
37
If you open the comments on this one itβs literally all spam. Insane. Same thing with my DMs. Itβs a major problem to my experience of the product β albeit honestly it doesnβt seem crazy hard to fix.
You canβt ban your way to number one.
3
2
38
Refactoring is recommended by many newsletters by now, but it hits differently when @SubstackInc itself recommends you! π₯. I woke up today and saw Refactoring up there, right below publications by the Substack staff itself!. Thank you so much guys for supporting my work πβ€οΈ
7
1
34
Introducing the new Refactoring π. Today I am introducing the most important changes to Refactoring since the launch of the paid plan!. They are the result of months of work and countless conversations with readers about how to make the newsletter better. Let's get started π.
3
7
35
Treat all relationships like long-term relationships. This is one of the most important lessons I have learned in my life. When we think long-term, we usually optimize for the right things. This is true in almost any area of life, but especially in dealing with other people.
2
4
33
In September 2020, after 8 years as co-founder and CTO of my startup, I quit to join a larger company. I went from a place where I knew all the ins and outs, to one that was completely foreign. It was hard. That's why you need a plan π.
2
1
31
How I Grow My Newsletter π. went from 0 to 12K subscribers in less than one year, without any previous audience to leverage (I had 100 twitter followers). Here is what I did (and still do) to make it grow. π§΅π.
7
4
33
The easiest way to make higher quality PRs is to make smaller PRs.
3
4
31
Consider theΒ Gall's Law:. A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works and cannot be patched up to make it work. You have to start over with a working simple system.
1
7
30
@shl Well, this is true as long as you need to scale ~2x. To scale 10x you need specialists and processes. Can't escape that. The other problem with "generalism" is you always need incredible people. My take is: first 10 hires = incredible generalists, then scale vertically.
0
4
28
βοΈ Play the long game. Despite popular narratives, I believe most success stories are not born out of incredible performance leading to overnight success. They are born out of adequate performance, sustained for an uncommonly long period of time.
4
3
29
Today's guest is none other than Charity Majors (@mipsytipsy), CTO at @honeycombio and my favorite writer. During our chat we talked about:. 1) βοΈΒ Observability vs monitoringΒ β whatβs the difference, and what good observability enables you to do. 2) π¬Β Intercom migration story
2
5
29
Yesterday we kicked off the podcast season with a fantastic chat with @lrnrd π₯ about managing teams in difficult times. The event ran live for the Refactoring community members (thanks to everybody who tuned in, left comments and asked questions π) and will go out next week
π₯ Introducing the Refactoring Podcast! π₯. Big newsss!! I am launching a brand new podcast for Refactoring!. Every week I will interview a world-class engineering leader about making software and working well together. Here is a peek at the first season π
0
0
4
New podcast episode out! ποΈπ₯. Todayβs guest is Kent Beck (@KentBeck), original signer of the Agile manifesto, creator of Extreme Programming, and all-round programming legend. With Kent we had the unique opportunity to talk about theΒ big picture. IΒ inquiredΒ him about the state
2
8
30
@shl I love brevity on Twitter, but sometimes I wonder if this kind of tweets really helps people?. Hard to understand + high chance of being misunderstood + little connection with actual work. Love your work but lately I find it hard to relate with your tweets π€·ββοΈ.
4
0
27
Bad product teams are given solutions to implement. Good product teams are given problems to solve. Great product teams are given agency to find problems and provide solutions.
1
5
29
One of the most influential works on structuring effective communication is The Pyramid Principle, by Barbara Minto. The principle advocates that ideas in writing should always form a pyramid under a single thought π
1
5
29
Minimum Viable Testing π. In the past few weeks I have interviewed tens of tech leaders about how they test software. I took a ton of notes and put everything together to design a 5-steps process that works, without overwhelming engineers. Here it is π
1
7
30
90% of a retrospective success is about psychological safety. Make people comfortable speaking their mind and the process will take care of itself.
1
5
27
π The Four Types of Work. As a CTO, for years I struggled with planning the work at my startup. I was young, and I eventually figured there are different types of work, and these should follow different processes. So I ended up writing a small framework about it. A thread π.
2
6
28
βοΈ Boring vs Exciting Tech. Choosing tech is hard. Pros and cons are usually: . β hard to articulate.β obscure to most decision makers. As a result, many choices end up being backed by little more than "gut" or personal taste. Is there a better way? Let's figure this out π.
1
7
25
π 1 Year of Refactoring. This week marks one year since I started writing Refactoring! Here are a few numbers:. βοΈ 56 articles written.π 173 resources linked (~3 per article).π₯ 11,824 subscribers.π¬ 138,123 email opens. A thread π.
5
1
27
One of the most underrated tactics to onboard yourself in a new company is to have 1:1s with literally everyone who is even remotely related to your job. Do that in the first 30 days, and don't wait for your manager to take care of this. How to?. 1) Create a list of all the.
2
5
28
@KentBeck I have changed opinion on this many times over the years. What I would do today is probably:.- normal changes (most) = non blocking review.- critical changes (few) = blocking review.- trivial changes (few) = no review.
1
2
26
π§βπ€βπ§ How to Run Effective 1:1s. There is a lot of talk about reducing meetings and going for async communication. Rightfully so. But there is at least one meeting we should keep and get right: 1:1s. A primer ππ§΅.
3
7
26
Refactoring just crossed 60,000 subscribers! π₯³. This is the fourth year of the newsletter, and the progression has been incredible:. 1) 2020 β 900 subscribers.2) 2021 β 10,000 subscribers.3) 2022 β 25,000 subscribers.4) 2023 β 60,000 subscribers. Usually, the end of the year for
7
1
27
"The output of a manager is the output of the teams under their supervision or influence." β Andy Grove. Easy to agree on, hard to come to terms with.
0
7
25
Writing good docs is tricky π. In my experience, most teams have either a good workspace for docs, where everyone contributes, or basically no reliable docs at all. There is hardly a middle ground. That's because docs naturally lead to either a virtuous or vicious cycle π
3
1
27
"Maker's schedule, Manager's schedule" is one of my favorite Paul Grahamβs essays. In it, he explains the radical difference between two types of work π
2
1
27
"Management is doing things right. Leadership is doing the right things.". β Peter Drucker.
0
4
23
@paulg I see but I am likewise skeptical of Elonβs tweets β he has an agenda, too. I guess thatβs what happens when things become very divisive: you just canβt trust any source anymore.
2
0
24
The best thing I did recently is to create a weekly digest of articles written by people I follow on twitter. I *love* receiving this email every week. I used and @mailbrew. Huge props to @frankdilo and @linuz90 for making it possible π.
0
4
25
The Agile manifesto is 20 years old now, but it is still very relevant. I believe there are four major themes we can take from it:. π₯ Work closely with stakeholders.π Work in small batches.π
Give teams agency.π¨ Promote simplicity and good design. I think at them often.
4
5
25
Consider the Cockburn and Williams research on pair programming. 80%+ of devs who do pair programming are happier at work *because* of it.
1
3
24
πΈ Technical Debt it's a great metaphor that everybody in tech understands. That's because over time, all engineering teams get slowed down somehow by the existing codebase. π€ But why? . A thread π.
1
10
24
Many companies are reluctant to create engineering principles because they believe they are abstract and in the end they are not that useful. This is false π.
1
1
25
@t3dotgg How can you do the comparison without knowing what a Basecamp request does? . Maybe they batch a lot of stuff in them and so they need more compute. We donβt know. Cmon Theo you are better than this.
2
0
25
Pair programming is best when it allows you to produce high quality software and share knowledge in the process. For this reason, asΒ @kentbeck says:. "Pairing works when there is sufficient uncertainty in the problem be solved and the approach to solving it.".
1
12
25
Staff meetings at Amazon begin withΒ 30 minutes of silent reading. The purpose and agenda of each meeting is written down in advance on a 6-pager. Participants read it silently at the beginning of the meeting.
4
6
23
A common fallacy is mistaking manager behavior for full-time management. You should set up basic manager behavior β e.g. 1:1s, hiring, project mgmt β much earlier than you appoint full-time managers.
1
3
23
I believe tech requirements for most engineering jobs today are too prescriptive. I usually keep them loose, for 2 reasons:. 1) Smart people in tech come from various backgrounds you can't anticipate.2) Smart people learn what they don't know faster than most of us realize. π.
4
2
21
Two weeks ago I sat down with @jamesacowling, former Senior Principal Engineer at Dropbox, and we had a great chat about various topics. James joined Dropbox when it had less than 100 employees, and saw it grow to more than 2000. A thread π.
1
4
22
Refactoring just crossed 40K free subs and 1200 paid subs! π₯³. This is another amazing milestone β thank you all for supporting this journey and trusting me with your time and attention! π
3
0
21
I love David Pink's theory about the three elements that drive our motivation:. π Autonomy β the desire to be self directed. π Mastery β the desire to improve our craft. ποΈ Purpose β the desire to have a positive impact. This applies both to teams and individuals.
2
5
22
Writing is networking for introverts. It makes like-minded people reach out to you, drawn by your writing. One of the internet's superpowers is how easy it is to connect with like-minded people. Writing really supercharges this.
3
3
22
One advantage startups usually have over big companies is they can hire people with unorthodox backgrounds. Self-taught, generalists, people with weird journeys β they often make for great hires, while are a tough sell for Google or Facebook.
5
5
21
I can tell you: with the rightpeople, any process works, and with the wrong people, no process works. Processes are important, but they are a means for standardization and enforcement. As such, they can only guarantee adequacy, not greatness. Greatness comes from people.
5
1
21
Honest recommendation: a few months ago I bought a standing desk and it genuinely changed my life. I now spend about half of my working hours standing, which also means I walk more around the house. I feel like I have way more energy and better posture. Very recommended.
2
0
21
One of the most underrated tactics to onboard yourself well in a new company is to have 1:1s with literally *everyone* who is even remotely related to your job. Do that in the first 30 days, and don't wait for your manager to take care of this. How to? π.
2
2
20
Itβs 2023 and I believe we should agree on one thing: velocity, as an engineering metric, is not very useful. In my experience, velocity has limited utility for planning work. You can use it for rough sizing, but nothing more. For anything else, it is just misleading. In fact:.
3
4
21
π© Feature flags β build vs buy βοΈ. FFs look like a simple system. In fact, many teams I know have built in-house solutions to support them. However, since flags are useful for so many purposes, the scope of these systems tends to *creep* over time π.
1
4
20
Some of the best-performing engineering teams DON'T use a Staging environment to test things before release. They just push to production. 1) Why?.2) How do they pull this off?.3) Should you use Staging yourself?. I talked with tens of teams and are the lessons I learned π.
4
3
20
I believe the reason why pair programming gets so little adoption in most companies is because it is wildly counterintuitive. In fact, it looks bad to all stakeholdersπ.
1
3
19
π Underperformance Checklist. I believe 80% of performance is systemic, rather than individual. So, with underperformers, it is useful to ask yourself questions about the *context* they are working in. Here is a 15-items checklist I built for this over time π
2
4
20
Engineering Principles β. Autonomy is one of the key qualities you should foster in your engineering team. It drives motivation and makes the company scale. There are three elements that enable autonomy:. π― Goals.π Skills.β Principles. A thread π§΅π.
1
1
19
Generalists π¨. When my experience as founder ended after 8yrs, I felt uncomfortable with most other roles I could apply for. β I felt unprepared: none of those had been my full-time job.β Didn't want to leave behind other skills I genuinely loved. I had become a generalist. π§΅π.
3
5
18
One of the strongest benefits of pair programming, that few realize, is that you don't need code reviews. Think about that β code is automatically reviewed by your partner. You can release it as soon as it's written. And your partner review is 10x better than your usual PR.
0
5
17
A modest proposal: let's focus on 10x teams instead of 10x engineers. Great teams create great engineers, while the opposite is not always true.
0
4
20
Mixed Seniority in Engineering Teams π©βπ¦. I believe mixing Junior and Senior devs is key to make people grow and create healthy team dynamics. This is somewhat counterintuitive. If I can afford them, why shouldnβt I just hire the best engineers I can?. A thread π§΅π.
1
3
20
In 2+ years of writing Refactoring, I have read 20+ books and hundreds of articles about good management. One of the biggest fallacies I see in this space is to think that we can take frameworks and processes and turn any team into a great one.
1
1
20
You go to the gym to build a healthier body. You write to build a healthier mind.
2
3
17