Julian (@jsch4efer) profile picture
Joined May 2020
Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)Photo by Julian (@jsch4efer)
Everyone seems to have forgotten that code is a liability, not an asset. That means that producing thousands of lines of code you didn’t read isn’t something to celebrate. It’s something to worry about! Every line of code needs to be tested, secured, and maintained. Even the ones that create no value for customers and therefore no value to the business. That’s why every line of code is a liability. And producing more liabilities isn’t good engineering or good business. Good engineering isn't about producing more code. It's about solving problems with the least amount of code. The asset is the solution, not the code.
154
278
39
2,127
62,146
Julian retweeted
Asked one of our developers how he uses his AI agents and he just sent me this:
179
1,337
270
16,199
1,779,780
Hard times create strong types Strong types create good times Good times create weak types Weak types create hard times
Replying to @ShriramKMurthi
Meanwhile the rest of us schmucks had to make do with this
4
39
3
297
6,638
Julian retweeted
My further (informed) fantasies about the near future: 1) No CI. Agents prove each change against rules while they write it. 2) Code is a cache: rules stay, code regenerated with better models. 3) No staging, simulation instead. 4) Humans direct in plain words, agents compile that into formal rules, humans don't read the formulas, only examples and counterexamples. 5) Agents find the rules the human never knew to express (and ask to confirm them). 6) The checker catches contradictions between human's rules while they're still directing. 7) Unchecked/unproved code becomes legacy code. 8) Branches come back as experiments (many versions at once, humans pick among winners). 9) Software gets bred, not built. 10) Slop is variation without selection. 11) If UI is generated on the fly, design becomes the rules every generated screen must follow. A checker runs those rules on every generation. 12) Apps evolve from how they're used. 13) Everyone uses their own software version. One product becomes many personal versions, that stay compatible via shared rules.
Next question: why are you still using branches? We talked about PRs. I don't think they make sense anymore, if you rethink them from first principles. Yes, if you only use them to trigger CI, then sure, go ahead. That's a fat commit, not a PR. But for mandatory code review? Knowledge sharing? C'mon. But some of you said: "we use PRs because we use branches! I love to keep my stuff on branches." And I'm here to tell you: you don't need branches! You can just push to main. It's called trunk-based development and the idea behind it is true continuous integration: you continuously integrate your changes with the trunk, the main branch. It's been around for a while, it works. Just yesterday I talked with someone who said they've used it in an org with 400 engineers, without any issues. Ruby, the project, uses trunk-based development. The main idea is that you want to avoid long-lived branches. They're bad, because branches can diverge and merging can turn into a thing. And you don't want routine operations to become a thing. Even when I worked in teams that used PRs and branches, I considered it a smell when a branch lived for longer than a day or two. But when I shared this with others, I'd often get something back like this: "Well, good for you, but no. MY thing, what I'm working on, that can't be done on main. It needs to be on a branch! It has a database migration, it changes the underlying shape, the architecture, it refactors this and changes that. It can't go on main yet! I need to merge it all at the end." And then I'd say: "Nah, I can merge nearly anything continuously in small commits to main. You split things up. Add the column first, merge that, deploy it, now it's out, works, safe. Add the helpers first, merge them, get them out. Want to migrate from legacy pattern A to new pattern B? Don't do it all at once. Add the dependency, then migrate A to B, one by one, safe." If it can be merged in a large change, I bet it can be done in small increments that you merge to main. Yes, there are exceptions, but it's 0.5% of cases, not 50%. But there was a real problem with this approach: it was hard! There's a lot of "meta programming" you have to do: make sure this small first change compiles and runs and the tests work, merge it, get it out, monitor it; then do the next small piece, make sure that works and get it out; then maybe push some still-unused dead code, add a feature flag, add a config variable; or keep both configurations working at the same time, etc. That meant sometimes it was just easier to do it in a big fat branch. But big fat branches are risky! A system that continuously flows is safer than a system that starts and stops in spurts. But -- you guessed it -- here's the good news: Now, with agents, all of these chores are no chores at all! Agents are great at adding feature flags, they're great at splitting up code, they're great at keeping track of three changes and merging them one by one and watching them roll out. They're great at keeping two configurations working for a day. They're great at migrating one by one. In other words: it's now 10x easier to just work on main and make sure everything is continuously integrated and working. It's now easy to make sure the system constantly flows. And look, ... I'm not here to convince you to switch to trunk-based development. Just a ruse to get you to read this. (But also, hey: try it! Pushing to main is amazing.) I'm here to make you RETHINK how you develop software. Don't just use agents to replace your typing. Use agents as the magic code-writing machines they are. Everything that was hard about software development because it involved writing code -- even the meta programming -- has become easier. You can now reach for the stars! But you need to look up.
33
24
12
399
31,547
It's Halloween season again and as a mathematician I'm obliged to post this. I don't make the rules
65
4,830
179
20,006
415,068
Replying to @miniapeur
4
47
3
1,111
56,810
We are about 3 months away from someone coming clean with this story: We spent $4m in tokens to port our 2.7m LoC enterprise Java application to Rust. - P99 performance increased by 0.3% - RAM usage decreased by 2% - Full compile times increased by 437x and required 279GB of disk. - Incremental compile times increased by 9853x and required 279GB of disk. - Average developer salary increased by 24% - No one understands the code - We are super hip again!!
93
122
13
2,352
149,790
Julian retweeted
This is great news for AI, formal methods and the industry. My background is in safety engineering, and I’ve spent much of the past two years writing and speaking about safety for software built by agents and for agents taking actions autonomously. It has often felt like a fairly small conversation, so seeing @bcherny discuss his use of Lean and TLA+ and @nvidia use formal verification in OpenShell get this attention over the past two weeks is refreshing and validating. It feels like a mainstream moment for this work. In my talks at @AnthropicAI and @AntithesisHQ (I'll link in comments), I’ve argued for higher-order abstractions and formal specification and verification methods, not just instructions in Markdown files. If we’re serious about being accountable for software and applications written entirely by agents, we need to be able to specify their behavior, verify it independently, and enforce those rules in the running system. The OpenShell research post (link in comments) has a very good introduction to SMT solvers and shows how Z3 can check whether a proposed agent policy permits anything outside an approved boundary. AWS’s automated reasoning team has been applying these methods to access policies for years. I’d recommend reading it even if you haven’t worked with these tools before. Our Systems Research team at @datadoghq are researching the composition of SMT solvers, alongside model checking, deterministic simulation, and property-based tests to check the behavioral specification of a cognitive application as a whole. We’re putting these methods together in a runtime (called Temper) where the specification is part of how the application runs, with permissions and state changes enforced by the system. There’s a lot of overlap in the thinking with openshell, even though the checks operate on different parts of the software. We will be integrating with and contributing upstream. I’m very happy to see NVIDIA bringing the safety work to a much broader audience, and hope it helps more organizations put time into the rigor and accountability of agent-built and agent-operated software. I’ll add a few of my earlier posts on this for anyone who wants to read more, and I’m excited to share more of what we’ve been learning. teapawt.net/nerdsane/status/210391…
Today, with over 100 industry partners, we introduced the NVIDIA Open Agent Safety Platform, bringing together OpenShell and Sentry. Artificial intelligence is extraordinary technology that will advance discovery, productivity, security, health, and prosperity for generations to come. But its full promise can only be realized when people have confidence that AI is being built to be safe and deployed with wisdom and responsibility. This is bigger than a single product. It's the beginning of an open ecosystem to build the trust layer for safe agent systems. Together, we are building the foundation of the AI economy. Trust and innovation are not in conflict. Safety is how trust is earned. We must build not only the most capable AI, but the most trusted AI, so that this extraordinary technology can realize its enormous promise for the world. nvda.ws/4hcoq7m
4
6
1
42
5,307
Julian retweeted
>be senior devops engineer in Germany >master's degree, 10 years experience, five interview rounds >72,000€ before tax >C2 German required >All team members based in Asia speaking English >mandatory office attendance to sit on Teams with people at another mandatory office >"we have a startup mindset" >company founded in 1894, CEO is the founder's great-grandson >"we have flat hierarchies" >eight people can veto your decision, none can approve it >hired to "drive digital transformation" >need department head's signature to install Docker >service account request rejected because "accounts must belong to a person" >CI/CD pipeline ends with a ticket for Jürgen >Jürgen on holiday until the 23rd >AI tools forbidden because Datenschutz >database password in an encrypted Excel file called Passwörter_final_NEU.xlsx >root password in Confluence >on-call every weekend >company phone listed under benefits >ask for a raise >unfortunately the salary band is exhausted >no budget for a second monitor >external consultant gets 1,400€ a day to ask you how the code works >take banana from company fruit basket >office manager says those are for the upcoming customer visit >Fachkräftemangel
441
1,010
148
17,131
936,662
Having the repository keeping track of its own (potential) evolution is a beautiful thing.
interesting concept. I’ve always loved TODO.md
3
55
Replying to @BullTheoryio
27
447
16
18,931
576,756
We're in the weirdest job market of all time. From 2020 to mid 2022, companies were hiring at a pace that made no sense. Some teams grew 50%, some doubled. Everyone was afraid of missing out on talent, so they just kept adding headcount. If you could spell the word "javascript" you could land a remote role and a 25% raise. Then the second half of 2022 hit and the hangover started. Layoff after layoff. Each wave was supposed to be "the last one." Now in 2026 there are already over 130,000 tech layoffs and we still have another 6 months to go in the year. The twist this time is AI. Companies aren't just saying they overhired anymore. They're saying AI is making them leaner. That they can do more with fewer people. It's become the convenient new reason that sounds strategic while making their stock pop. But here's where it gets absurd. More and more companies are admitting the AI math isn't working out. They can't find the ROI they promised their boards. The computing costs are massive. It's more expensive than they thought, not less. So let me get this straight. We're in a job market where companies fired their workforce to buy something they now say is too expensive and doesn't work as advertised. And the executives responsible for those decisions? They're still collecting their bonuses. Weirdest job market of all time.
123
248
40
1,946
271,061
This is going to be long. Last semester I suspected I had a major issue with use of AI in my survey courses, so I inserted what is known as a trojan horse (not the virus kind) into the directions of a paper assignment. As it turned out, I did in fact have a major problem, and a post on Threads about it accidentally went quasi-viral and ultimately became a Huffington Post article and an NPR interview. (Links at the end)
24
201
55
1,650
260,701
Claude Code is Farmville for 40 year old former software engineers
133
379
91
5,696
597,076
Semantic web revival incoming 🤗
discovering basic web dev from first principals
44
> In this new world, high quality software is infinitesimally available. 🤣
> In this new world, high quality software is infinitely available. Hahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahahaha
66
Well, I guess I'm 40 now 🥲
Emacs IS an age verification scheme. Nobody under 40 uses it.
1
62
* and the ability to express it.
taste is a new core skill
50
Last dev activities before a customer demo 🤣
I can’t tell if this guy is an Olympian or if someone just pushed him down the hill?
310