There’s no occasion, no special reason for why I’m putting this in here, other than an uncontrollable feeling that I have to.
My incomplete list of advice for How to Demo (In Case You Need to Demo a Feature or a Bug Fix That You Worked On to Your Colleagues in the Form of a Video or a Short Presentation over Screen Share):
Tell the audience what problem is being fixed; tell them why you did what you did.
Don’t interrupt yourself.
Don’t let the presentation and your talking go out of sync.
Keep it short.
No, really, keep it shorter.
Call out what's important! People won't notice that it's now a 4000 line file on which you demonstrate your whitespace-analysis tool, you have to call it out and say "watch, this even works on a 4000 line file"
Do not show unnecessary stuff. Don't go fullscreen-recording if you want to only show your terminal. People will get distracted by your battery levels. They will ask themselves "what is this icon in the menu bar?" instead of listening to you.
You don't have a laser pointer and you don't have an index finger — make up for it. Wiggle your cursor, make a selection, select some text. People can't see what your eyes are focusing on. Show them what they should focus on.
Every frame, every scene: know where the user should look.
Use 1080p.
Bump that font size. And once more.
Don't say "it's probably not gonna work" when demo doesn't aim to show how much of a prototype it is. Because if it then works despite your warning, your audience might miss it — you primed them to focus on things where it won’t work.
Pictures are worth a thousand words. You reduced logs the logs ingested by the system by 100x? Don’t just say that. Show the dashboard before and after. Show disk usage before and after.
And with that said: here’s a bag of links.
I was a guest on The Changelog podcast and the episode came out this week: Agent, take the wheel. What a title! I love it. Possibly even better, the teaser paragraph: “Thorsten Ball returned to Sourcegraph to work on Amp because he believes being able to talk to an alien intelligence that edits your code changes everything.” I mean, when you put it like that…? It was lovely talking to Adam and Jerod. My personal highlight and possibly a high point in my — well, let’s not call it career — work as a writer: Adam loved the original copy I wrote for ampcode.com and actually opened the archived version to read it out loud, in its entirety.
Thanks to this episode of Founders I ended up reading the 1985 Playboy interview with “Steven” (!) Jobs (here’s a more readable version.) I had never read it before, but while reading, I realized that this must be the original source for many of the Steve Jobs anecdotes I had heard. It’s all in there: the story about how he got his summer job at Hewlett-Packard (“When I was 12 or 13, I wanted to build something and I needed some parts, so I picked up the phone and called Bill Hewlett—he was listed in the Palo Alto phone book. He answered the phone and he was real nice. He chatted with me for, like, 20 minutes. He didn’t know me at all, but he ended up giving me some parts and he got me a job that summer working at Hewlett-Packard”), mentions of Shakespear and Bob Dylan along with declaration to inject liberal arts into business (“There’s always been this myth that really neat, fun people at home all of a sudden have to become very dull and boring when they come to work. It’s simply not true. If we can inject that liberal-arts spirit into the very serious realm of business, I think it will be a worthwhile contribution.”), California, LSD, the love for computers (“A computer is the most incredible tool we’ve ever seen. It can be a writing tool, a communications center, a supercalculator, a planner, a filer and an artistic instrument all in one, just by being given new instructions, or software, to work from. There are no other tools that have the power and versatility of a computer.”) Remarkable interview, highly recommend reading it.
“Steve Jobs gave a great talk at MIT in the 80's about Operational vs Managerial”, but the link points to a 6min video. Really interesting (as is the original thread, which opened up a new world to me.)
For years and years and years (to be precise: five) I’ve told everyone who would listen that CleanShot X is one of the best macOS applications I’ve ever used. Last week someone asked on our work Slack: “can someone recommend a screenshot tool?” Tragically, I was asleep when the question was asked so couldn’t immediately say “CleanShot X, go and get it” but only six hours later, after others had already recommended it, which made me look like a pile-on kinda guy instead of the true, hardcore, before-it-was-cool
fanconnoisseur that I am. Put in the wrong corner, I now had to fight my way out of it, so I recorded a 5min video in which I explained why our Head of Engineering, Erika, should buy CleanShot X and posted it in that Slack thread. Four colleagues bought it because of that video, so I posted it on Twitter (don’t watch it if you don’t want to fall in love with a piece of software) where it got 233 likes and caused to the co-founder of CleanShot X to reach out to me and offer me the chance to become something I’ve never been in my life: an affiliate. So, with the warning that this will trigger all of the adblockers, here it is: my affiliate link with which you can buy CleanShot X and give me some of that money as a token of appreciation for writing this much-too-long-now paragraph and recording that video.The New Skill in AI is Not Prompting, It's Context Engineering. Yup, but I wouldn’t say it’s new. It’s been the game the whole time.
Drew Breunig wrote two really, really, really good posts on context: How Long Contexts Fail and How to Fix Your Context. I’ve already quoted them multiple times this week and one has given me a neat idea I want to build into Amp.
Figma filed an S-1 in preparation for an IPO and you can actually read the whole thing here, on the sec.gov website, and it has some beautifully designed pages next to some very official looking tables and the experience of just scrolling through the whole thing is very interesting. I haven’t read it all (I’d love to be the type of person who reads S-1s, but alas), just the opening page (“When Evan and I started Figma in 2012”) and the fourth paragraph contains this: “To be clear, that's not a promise of share price growth. Even if we execute perfectly (we won't — no one ever does) markets wax and wane. In addition, you should know that while we've built an efficient business, our primary goal is not efficiency. Our goal is to achieve long-term growth by supporting the rapidly evolving needs of designers.” That’s right, in the S-1.
Talking about Figma: if you haven’t, I highly recommend reading through this page on Figma, created by Figma’s co-founder, Evan Wallace, describing his contributions. And then you also need to know that he created esbuild.
André Arko: You should delete tests. Yes! Amen. Tests fulfill different jobs: some help you discover an implementation, some help you nail down behavior, some help you stop behavior from changing. And sometimes the tests have done their job and it’s over and they’re done and you can delete them. Some tests are just a ladder on which you climb up and once you’re up, you can remove them. The sacred status that some test code has achieved in some codebases is mind-boggling. It’s just code. It can be a liability, too. Ask yourself: does this test code still do its job? Or will it never, ever fail again and is now just a relict from a time in which we didn’t know how to write the implementation we have now? Delete it.
More people should write: “More people should do what I’m doing right now. They should sit at their computers and bat the cursor around — write full sentences about themselves and the things they care about. I have a selfish reason for my demand: I have a lot of friends who are thoughtful, but keep their thoughts to themselves. I imagine finding notebooks under their bed, tens of composition books packed with little print. I think about what sort of a treasure that would be. But that’s not why you should write.”
Lovely, lovely, lovely: macOS Icon History.
“The best statistical graphic ever drawn” unites “six different sets of data”: geography, the army’s course, the army’s direction, the number of soldiers remaining (“each millimetre represents 10.000 men”), temperature, time. Wikipedia has a version you can actually look at.
I read this little book in two evenings: David Shield’s How We Got Here. I wouldn’t say that I… loved it? Or that I even know what it wants to say, but, to quote this recent Economist bit on Adam Curtis, who I thought of multiple times while reading the book: “It is strange stuff, even with the help of explanatory subtitles, but the cumulative effect is compelling.” If you have a spare evening or two and want to go “huh?” — there you go.
This is great: I Shipped a macOS App Built Entirely by Claude Code. Full of practical advice and nicely formatted and screenshots and … it’s great. (Obviously this could’ve been written about Amp too: I’m pretty sure it could’ve done everything mentioned here and then some.)
Matthew Prince, Cloudflare’s CEO: “The future of the web is going to be people reading the summaries of content, not the original. What I'm worried about is - if you can't sell subscriptions or monetize ads or get the ego boost from people reading your stuff, why anyone is going to create content?”
“I want you to pour your heart and soul. I want you to rewrite. I want you to revise. I want you to make something that you actually care for, that’s like a risk, a risky expression of who you are, what you stand for, what you believe. That’s what I want to see.” (I feel a little bit dirty linking to a YouTube Shorts video. Or do we say a short? In any case: I ain’t judging. Go and watch Billions in 60 second clips.)
“Sometimes the process, the checklist, the logic — it blinds us to what’s right in front of us. Because what’s right doesn’t always fit. It doesn’t always come in the expected form. And it almost never arrives at the perfect moment.” Articles like this do feel like the YouTube Shorts of writing, but this one does have Michael Jordan in it and ever since watching The Last Dance I’m a sucker for anything with Michael Jordan in it.
The Benchmarking Paradox: “Those deeply familiar with a given system to tune it right, may be biased to favor it in comparisons, whereas those trying to create fair comparisons oftentimes lack the deep insight required to tune all the systems right.”
This article is great and funny and nuanced and goes deeper and, now, in 2025, has more dimensions than you might think after reading this very strong opening paragraph: “The new media technology was going to make us stupid, to reduce all human interaction to a sales pitch. It was going to corrode our minds, degrade communication, and waste our time. Its sudden rise and rapid spread through business, government, and education augured nothing less than “the end of reason,” as one famous artist put it, for better or for worse. In the end, it would even get blamed for the live-broadcast deaths of seven Americans on national television. The year was 2003, and Americans were freaking out about the world-altering risks of … Microsoft PowerPoint.”
Marc Brooker, if he “could offer you a single piece of career advice”: “avoid negativity echo chambers. Every organization and industry has watering holes where the whiners hang out. The cynical. The jaded. These spots feel attractive. Everybody has something they can complain about, and complaining is fun. These places are inviting and inclusive: as long as you’re whining, or complaining, or cynical, you’re in. If you’re positive, optimistic, or ambitious, you’re out. Avoid these places.”
We are looking to hire a Senior Design Engineer to work with us on Amp. This week, I said to a friend: “What I really need right now is people that know when it's good for a project to do things ‘like they do them at Google’ and when to play startup-land dirty and how to switch between both modes multiple times a day.” If that’s you, in code and design, I’d love to work with you. (Since this is my newsletter and I can do with it what I want and because I know that most, if not all, of you are talented, smart, hard-working, beautiful, funny, and charming, I thought: there ain’t a better place to reach out.)
This week I was reminded of John Carmack’s exit note when he left Meta: "We have a ridiculous amount of people and resources, but we constantly self-sabotage and squander effort. There is no way to sugar coat this; I think our organization is operating at half the effectiveness that would make me happy. Some may scoff and contend we are doing just fine, but others will laugh and say 'Half? Ha! I'm at quarter efficiency!'" What I constantly — nearly every week — think of is the last line:
”Make better decisions and fill your products with ‘Give a Damn’!”
