Galo Navarro

Code reviews are not sustainable

@GergelyOrosz Or residence in non overlapping timezones!

Devs who can code also WITHOUT AI as well looking to became 10x more valuable They are the ones who won’t panic or be idle when their Claude quota runs out… So much for all the advice on how learning to code is not worth it any more…

@GergelyOrosz Funny how we keep rediscovering Fast, Cheap, Good, pick two.

... or sacrifice quality. Or performance. Or maintainability. Or anything that us devs would call "non-functional requirements." The stuff that is invisible on first glance... until it's not, and makes the difference between a product users love, and one they hate

To this point: the litellm supply chain attack calls for re-thinking the tradeoffs around automated dependency updates. Any decent sized project has a constant stream of dependabot updates, which are impossible to review thoroughly, and are a vector for these attacks

I honestly think the entire AI wrapper discourse is backwards. People keep saying "it's just a wrapper" like that's an insult. Gmail is a wrapper around SMTP. Stripe is a wrapper around bank APIs. The wrapper IS the product. Always has been. The model is the commodity.

@artman Oh in case it helps the GHCR mirrors work well as they are not on cloudflare

Hasn’t hit me yet, but I dread the inevitable when we have an outage while a football match is going on and I waste an hour figuring out why I can’t access some of our services.

@artman It hit me for the first time a couple of weeks ago, VPN deployed and ready to use

@yuzu_4ever “Si tu no tienes felicidad, de sabio no tienes na” ~ “If you ain’t got happiness, you ain’t got wisdom”.

i am always suspicious of people who claim to be smart but are miserable. if they really were that smart, they would have figured out how to live a joyous life.

@mitsuhiko Have you tried Hegel?

Having an agent edit its own edit tool like and debug it is the most meta experience possible.

@ayushagarwal From varoa.net/2026/01/30/poc…

the worst lie in tech is "I'll add tests later." no you won't. none of us will. that code is going to production naked and afraid.

@loujaybee I am thinking of it as airport security with layers of checks that flag changes based on heuristics and divert from automated gates (linters, tests..) towards human ones (reviews, judgment calls..)

For those of you experimenting with software factories, what are you using as your 'co-ordination' layer? How are you gating steps in your workflow?

From brooker.co.za/blog/2026/03/2…

I'm not very happy with the code quality and I think agents bloat abstractions, have poor code aesthetics, are very prone to copy pasting code blocks and it's a mess, but at this point I stopped fighting it too hard and just moved on. The agents do not listen to my instructions in the AGENTS.md files. E.g. just as one example, no matter how many times I say something like: "Every line of code should do exactly one thing and use intermediate variables as a form of documentation" They will still "multitask" and create complex constructs where one line of code calls 2 functions and then indexes an array with the result. I think in principle I could use hooks or slash commands to clean this up but at some point just a shrug is easier. Yes I think LLM as a judge for soft rewards is in principle and long term slightly problematic (due to goodharting concerns), but in practice and for now I don't think we've picked the low hanging fruit yet here.

@eatonphil Xanax

what's everyone using for python dependencies outside of pip and uv?

You should probably enable TLS connection re-use/keepalive in your applications Here's a heatmap of tls.connect latency in a production Node.js app P50 is 400ms, P99 up at 3.9s Fix this and all your external API requests become instantly faster

students get 28% higher score simply for using pen and paper. it’s crazy. knowledge is not stored linearly in your brain, but as a graph of interconnected concepts. handwriting is more intuitive and engages your brain to absorb information than record it. it just works.

@clare_liguori GH would have built this by now if they had a serious interest but I think they won’t because it challenges way too many aspects of their platform

I work mostly in GitHub these days, but this one thing really makes me miss the internal Amazon code review system

When are we going to get serious tools for stacked pull requests? I generate a TON of code agentically, but then I want to split up the changes into bite-size chunks that can be reviewed by humans across multiple PEs (yes, I still think humans should look at code). My agent splits up the changes into multiple commits, but then what?

@ibuildthecloud The only animal that trips twice over the same stone..

One thing we've learn from mimicing the engineering orgs of big tech companies over the past 20 years is that you aren't a big tech company and they're solutions don't work for you, let alone them.

@adamhjk “Know thyself” was inscribed upon the Temple of Apollo in Ancient Greece. There is substance to the idea that the relationship with the self has changed over history though, but not sure that’s what he had in mind there

Yeah, because nobody ever introspected before 1910. Tell me you've never read, I don't know, the bible for $500. You think the guy who wrote the Psalms wasn't introspective? There is a vast difference between self flagellating guilt and the ability to evolve from past experiences through introspection.

Now that gitflow is finally buried is it time to rant about the NOJIRA nonsense?