Observability plumber @github. For data stuff, see @nora_neu. Post Musk society: @terrorobe@hachyderm.io Curated mastodon experience: @terrorobe.bsky.social
Gmunden, Österreich
Joined July 2007
- Tweets40.2K
- Following1.6K
- Followers1.4K
- Likes25.6K
Michael Renner retweeted
A conversation that has happened many times in banking, with various flavor texts, from fintech to public banking to microfinance.
A: We think there is an opportunity to write loans to an underserved population.
B: At usuriously high rates?
A: No, at usual rates. Maybe lower.
Michael Renner retweeted
In my career, compilers have always generated better code that I ever could have. I find this to be increasingly so with contemporary AIs.
But you know what neither can do better than me?
Knowing what software to write, what shape it should take, and what I want it to do.
The entire history of software engineering has been one of rising levels of abstraction.
This is how it was, how it is now, and how it shall be in the fullness of time.
Michael Renner retweeted
We made this prediction in a video of ours.
TLDR:
1) A few jobs are worth a lot (taste, decisions, deep-knowledge) and a lot of others are a race to the bottom where you compete with outsourcing hubs using AI.
2) AI will lead to a cambrian explosion of niche solutions as you can now service them with small teams
Full video on the channel: youtube.com/watch?v=zKeRUdnc…
Since the launch of ChatGPT, the Philippines offshore industry (8% of country’s GDP) has actually grown much bigger.
Employment in IT and business outsourcing is up by +20% to 1.9m workers and industry revenue is up +30% to $42B.
“AI is helping offshore workers in the Philippines do new, more complex jobs. They are picking up work training AI models or supervising AI agents. Hospitals in America are increasingly outsourcing the checking of insurance eligibility and filing of medical records to Filipinos packing AI tools. Mr Gallimore says more “high-value” work, such as accountancy or engineering, is going offshore in part because “AI is a leveller. You can now teach someone complex stuff quickly” and still get it done more cheaply than in America or Europe.”
***
More here: economist.com/asia/2026/08/0…
Michael Renner retweeted
"Everyone has everything now. All war is symmetric. Good luck with that."
Seems the jamming-resistant CRPA antenna on the Iranian Reentry Vehicle is a Chinese SWT-JRL1 according to @GEOIMINT
Michael Renner retweeted
We've recently made Pi's cache behavior more visible. This site has been debating whether agent harnesses are helping or quietly torching their caches. That seemed like a good excuse to explain how KV caches actually work and how Pi helps (or doesn't). earendil.com/posts/prompt-ca…
Michael Renner retweeted
Wir kaufen die letzten 2 Monate jetzt aktiv und auf Masse Elektroautos ein.
Die Kisten stehen alle nicht länger wie 2-3 Wochen. Genug nachkaufen geht gar nicht 🤯😃
Michael Renner retweeted
The problem with the "if it works who cares what the code looks like" mindset for agentic work is that it assumes the agent has a perfect understanding of "works." Realistically, things are underspecified, agents make bad assumptions, etc.
To be fair, agents are pretty good at unit test coverage. They're pretty bad at designing human experiences (API, CLI flags, etc.), especially cohesive ones for future roadmap plans they may not have visibility into (unless your backlog is perfect and vision fully laid out, which I doubt). They're bad at knowing where performance matters and what type (CPU vs memory tradeoffs). They're bad at where compatibility matters and where it doesn't (and tend to err on the side of preserving it without further guidance). Etc.
Unless you have this ALL specified, you can't possibly claim "it works" without taking a look and thinking about it.
Michael Renner retweeted
I think I finally found the best way to interact with agents. Stole some parts from @badlogicgames pibot, mashed it together with @kyutai_labs stt and can finally be the man I've always wanted to be! Fully local on an m4 macbook btw!
Michael Renner retweeted
In case you are in the camp of “Andrew Tridgell is vibefucking rsync” please read this.l and adjust your priors. medium.com/@tridge60/rsync-a…
Michael Renner retweeted
(I have a complicated relationship with heroics, and to coin a phrase, the optimal level of heroics is not zero.)
Michael Renner retweeted
Ab 1. April gibt es reduzierte Netznutzungsentgelte #SNAP. Dafür muss man aber die Viertelstundenmessung aktiviert haben. Im Smart-Meter-Portal der Wiener Netze geht das unter Anlagendaten und "Viertelstunden-Intervall". Bei Netz OÖ zB kann man auch nachschauen, ob aktiv.
Michael Renner retweeted
A glimpse of the absolute insanity in numbers that is our industry right now. Attributed Claude activity over the last 90 days on GitHub. The pushers did us dirty over Christmas.
Michael Renner retweeted
Which local models can actually handle tool calling?
I built a framework to find out.
15 scenarios. 12 tools. Mocked responses. Temperature 0. No cherry-picking.
Tested every Qwen3.5 size from 0.8B to 397B, and since some of you asked after the distillation tests: yes, I included Jackrong's Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled too.
Only two models went all green: the 27B dense and the distilled 27B.
The 397B? Failed two tests. The 122B? Failed one. The 35B? Failed two.
The timed-out results — mostly on the smaller models, are cases where the model got stuck in a loop, repeating the same tool call until it hit the 30-second limit.
The test that exposed the most models: "Search for Iceland's population, then calculate 2% of it." Simple, but 35B, 122B, and 397B all used a rounded number from memory instead of the actual search result. They didn't trust their own tool output.
Small models hallucinate data.
Big models ignore data.
The 27B just threaded it through.
Michael Renner retweeted
recommended reading sure to ruffle some feathers. but it's largely true for now.
keeping the complexity off the bay is really hard, espwcially if you go full agent orchestration.
even if you don't, and human in the loop a lot, automation bias kicks in and your reviews of agent generated code become mostly performative.
