I'm a software developer and an SRE github.com/drogus/
Joined May 2007
- Tweets 7,981
- Following 703
- Followers 1,337
- Likes 8,059
166 Photos and videos
Jun 7
This might be harsh, but I really wish most of the infra software was not written in Go, but in a language that is easier to interface with, like Rust or Zig. I often stumble upon a good library or a program written in Go that is pretty much unusable from outside Go’s ecosystem
1
1
142
Jun 7
In the US they will show you “calories per serving”, and the serving size is like one chip
2
97
Jun 4
This may work if their output is already trash, so with multiple agents they can produce 100x trash output, no decision making is necessary if you go with auto mode 😆
Jun 3
I am so thoroughly convinced that anyone who thinks AI 100x's their output is a liar or a lunatic.
You are telling me you can make 1 years worth of decisions in 3.65 days? Let alone describing those accurately and coaxing the result from the AI... (1.8 days european time)?
1
113
Jun 3
People from US: stop treating Europe like it’s one homogeneous country challenge
33
Jun 2
It would break their brains if they learned that in many countries it’s mandated by law to let people merge in this situation
You’re standing in a long line at the deli. A guy walks in behind you, shifts to your left to walk next to the line and slides in front of the guy several people ahead of you. An urban planner reminds you that really is the most efficient way of managing a line.
95
Jun 2
“Soon we will have FSD”
No, we won’t
Jun 2
Vienna made it illegal to buy new cabs if they are not electric
So if you are a taxi driver and want to upgrade, you can’t buy a hybrid, diesel or a usual gas-powered car any more
It has to be either an electric vehicle or a hydrogen car
Lithium batteries are still unprocessable toxic waste we have no idea how to dispose off
So I think this policy will do more harm than good
It’s also weird to target a group of people whose livelihood depends on cars and make them implement your eco policy
Hybrid cars seem like the best eco option now so it’s x4 weird they are not included in the program
But then again, soon we will have FSD and taxi drivers will not be needed anymore (finally)
110
Jun 1
Deadlifting is not the problem here, one rep max tries are. Unless you are training for a competition, doing extremely small volumes, like 1-3, reps is pretty much useless, other than boosting your ego
May 29
my biggest regret in life is deadlifting.
i herniated a disc pulling 200kg. i had done it before, but that day felt off and i still went for it like an idiot. by 21, i was basically crippled. 8 years of rehab later and i’m maybe back to 40% of the back strength i had before.
if you deadlift, leave your ego at the door. low weight, more volume, clean reps. once you mess your back up, there is no glory. just regret.
1
89
May 28
That's some kind of cowboy coder and not a Staff Engineer 🤷♂️
May 27
a staff engineer at my old company got laid off during “cost cutting.”
his entire farewell meeting was 12 minutes long.
week later:
payment service started randomly failing.
turns out he was manually fixing edge-case data corruption every night for 3 years.
nobody even knew.
the most dangerous systems are the ones surviving because of one invisible engineer.
1
83
Piotr Sarnacki retweeted
May 26
recommended reading. i too am very done with people anthropomorphizing a bunch of matrices on a GPU cluster, especially if the same people do not give two fucks about actual human beings.
May 26
More musings after some people got upset about the word clanker. lucumr.pocoo.org/2026/5/26/c…
18
32
426
49,580
May 26
While I'm still not sure if LLMs speed me up when working on stuff that has to be production-grade (as opposed to personal projects), I know that I create much more tools - scripts, visualisation tools, debugging tools. Stuff that would have been hard to justify in the past
1
67
May 22
Tigerbeetle team is great, but this post is clearly in relation to Bun's Zig to Rust rewrite, and I don't think it's easy to compare projects with such different feature sets. The fact that Tigerbeetle can allocate all memory at startup makes memory management so much easier
May 11
We’ve had on the order of 3 memory bugs in 6 years of TigerBeetle. None RCEs.
On the other hand, our own simulators have proactively found hundreds of (devastatingly catastrophic) distributed systems correctness bugs per year.
Given how hard TigerBeetle’s domain is, in terms of mission critical financial transaction processing, I’ve never for one minute believed that writing TB in a memory safe language such as say TypeScript would somehow magically (!) make any material impact compared to the 100x correctness multiplier of TigerStyle.
That’s because—rather than fall for the fallacy of composition, i.e. to see distributed correctness as a language problem—TigerStyle instead takes ultimate responsibility for the “end to end” correctness of the distributed system as a whole.
Per systems engineering, correctness is always a systems design problem. For example, how to build a reliable whole, (especially) out of unreliable parts, such as broken firmware, bitrot, programmer error etc. In other words, application of the end to end principle.
But when you TigerStyle the design in this way, the world of systems engineering also completely opens up to you and changes how you evaluate systems languages (now things like “power to grammar ratio”, or explicitness, checked arithmetic and precision become more critical and valuable to you).
Of course, it is harder to care about correctness, to take responsibility for correctness end to end.
Yes, you’re forced to begin to worry about the more serious concerns, starting with the basics of static allocation, explicit limits, assertions, deterministic simulation testing and moving to more advanced topics like protocol-awareness and storage fault-tolerance.
But then again, TigerStyle is such a force multiplier, that you achieve mission critical quality, and in less time and with greater velocity.
If you’re tired of production issues, and if you want to “engineer your engineering”, I would encourage you to lift up your thinking to the level of systems design and end to end correctness.
Start thinking about your methodology and begin embracing TigerStyle.
tigerstyle.dev
6
64
26,133
May 22
Could Bun's team do a better job at using Zig in a way they don't have to constantly fight segfaults and other memory issues? Maybe, but interfacing with several other libraries is already a lot different than a self-contained database.
1
3
877
May 22
Imagine having to specify how much memory your script will use, up-front. Sure, with virtual memory mapping, you can get more than you need, but then releasing the physical memory gets tricky, I'm not even sure you can do it with Zig's allocator, and then you still risk UB.
2
738
May 22
I wish Rust incorporated de facto standard Rust libraries into stdlib. There is a set of broadly used stable libs that don't need to change fast nor massive API changes, so a slower stdlib changes cadence wouldn't be a problem, and it would really help with minimizing deps number
May 20
Fork your dependencies, trim them to only your use case, never update unless it breaks for your users. I’ve been vocal about this for 10 years. I’ve always said that updating is way riskier than latent bugs (which can be tracked and CVEs monitored).
If you are updating a dependency, it’s on you to analyze every single commit in the full transitive set of dependencies. If you dont see anything compelling, dont update!
I remember at HashiCorp once in awhile an engineer would try to update a dep or replace a DIY lib with an external one and id always ask “show me the commit we need.” Dont update for the sake of it.
Feeling pretty swell about this mentality with all the supply chain attacks happening.
1
249
May 16
No, APIs won’t be dead, and no, people will not expose “prompt as an API”, or at least not at scale.
May 15
RESTful APIs may be dead soon. Instead, web services may expose a single POST entry point for a prompt. Internally, an AI agent may decide how to interpret it and what to do with the data and the database.
1
131
Piotr Sarnacki retweeted
May 14
In March, we released benchmark numbers alongside SpacetimeDB 2.0. After a thorough investigation, we discovered a serious error in SpacetimeDB which caused our results to be misleading. We would like to sincerely apologize to our community.
Read the full article: spacetimedb.com/blog/benchma…
29
74
195
1,245,082
May 10
An interesting thing I’ve noticed lately is that I see more and more people use Rust (mostly vs Python) because cost of writing Rust code is much closer to Python with LLMs being able to figure out compilation/lifetime issues
2
27
2,734
Apr 17
I know that this is a weird hypothetical bait question, but I question anyone whose choice is Go. If the question was “which lang do you like the most”, sure, but having to build everything in Go limits what you can build
Apr 17
You can only build software in one for the rest of your life:
Go
Zig
Rust
What are you choosing?
226
Apr 7
After recent Claude vs OpenClaw weirdness, I'm wondering if this is on purpose, response from Claude when asking about OpenClaw: I don't recognize "OpenClaw" — could be a typo.
1
103
Apr 4
The problem I have with “just give an AI agent a good feedback loop” is youcan’t account for everything. How do you write tests that disallow security vulnerabilities?
94