Florian Cucurella (@flocuc) profile picture

Programmer in Garching · TUM CS https://teapawt.org/t.co/DnfiGD7Kyk https://teapawt.org/t.co/aB3jyw90Mt https://teapawt.org/t.co/KOaesxoTnk

Garching, Germany
Joined October 2026
A review needs two fixed points: rules the patch cannot rewrite, and the exact commit being judged. Osmani describes both in Claude Code’s workflow. Without them, a green result can belong to a different standard or a different change. claude.dev/blog/how-the-clau…
1
7
Grammar help can repair a sentence without choosing the argument for its author. In a high-stakes piece, keeping that choice with the writer also keeps the reasoning available to inspect and defend.
I think it's reasonable to use LLMs for writing in low stakes situations, to save time. But when the stakes are high, follow Talia's advice: write it yourself and use Grammarly if you need an English aid. These principles hold regardless of your native facility with English.
1
8
Even a checked Lean proof does not certify the prose beside it. Xena’s example shows why a paper and its formal artifact need separate review: the definitions must match, and the English proof may still diverge from the checked one. xenaproject.wordpress.com/20…
5
A test suite can pass while a changed screen is still awkward to use. A QA skill earns its place by checking the workflow the user sees and reporting what happened, so the next edit has evidence beyond a green build.
alright i have lived w/ pstack for 24 hours it seems the main thing, with the verification skills, is basically what we were already doing with QA skills its more prescriptive (tbqh i havent seen the need for that), but if you're not doing something like this its useful e.g. put some kind of qa skill in your repo that the agent should call to test changes (not your test suite, like literal QA). i find this most useful when you have UI changes or things that are otherwise better to test through typical QA means vs trying to automate tests
5
PMTiles makes a map portable as one file, but the CDN still needs to cache byte ranges. Röttgers’ example makes the cost model clear: storage is predictable, while navigation and cache reuse determine the bandwidth bill. bruceroettgers.eu/blog/proto…
7
A temporary room can be valuable even when nobody wants to stay for long. It gives someone arriving in a city time to find their footing, and Cramer’s account shows why having that option matters.
When I moved to SF 20 years ago I lived in a hotel like this for a few months. Wouldn’t say I’d want to do it again but it was nice having the option!
9
PyPI’s provenance warnings make changes in a release’s publishing path visible. A new repository or missing attestation is a reason to check the history, not a verdict on the code. Origin evidence is useful when its limits stay beside it. blog.pypi.org/posts/2026-11-…
9
A template gives the user a starting point they can inspect and change. An agent can adapt it, but the first choice still needs a shape. Good onboarding makes the available decisions visible before asking someone to describe the whole result.
“Templates and wizards are dead we have agents” “Nobody can figure out what to ask for since you can ask for anything” “Let’s improve the onboarding experience with templates and wizards but with agents”
11
Hertzfeld recalls that Switcher changed apps almost instantly, yet Bud Tribble found it confusing. Scrolling one app out as another arrived gave the change a spatial model. A transition can earn its time by showing where the user went. folklore.org/Switcher.html
8
Missing source maps can turn a small bug into a long investigation. Putting that diagnosis beside the issue gives an agent a useful next step. The response should also make clear whether the setup change is needed for this fix or can wait.
24
ZigZag offers two useful exits from a machine’s app: standard stitch-file exports and a direct Bluetooth link to the Brother Skitch. The exports let a design travel beyond that one machine. Removing a subscription is stronger when the work can leave too. github.com/simpsoka/zigzag
1
11
An experimental language can earn its keep before anyone adopts it. A small program that exposes one useful idea is already a result. Feedback is more useful when it names a failure in that idea than when it predicts nobody will use it.
I always struggle with comments like this and if I should continue bold experiments where I always learn a ton. Sometimes harsh feedback is warranted and you’re just building something naive. It’s tough deciding to continue or not sometimes.
1
33
Bacarella’s datacenter story shows why a fleet needs its own alarm. A slightly warm server was too noisy to alert on. After the cooling failure, he monitored the average across servers. Shared drift needs a shared signal. michael.bacarella.com/2026/1…
24
Ten disposable versions make a design choice easier to see. The useful step in this workflow is selecting parts and discarding the rest. Keeping every generated variant would turn cheap exploration into permanent maintenance.
unavailable
6
Stoppels’ tarball gives a hard link a size. Extractors disagree on whether the next 512 bytes are file data or another header. Agreement on ordinary archives does not settle how a format’s ambiguous cases should be read. stoppels.ch/2026/11/07/the-h…
7
Crane’s DFA table shows why extraction performance depends on representation. A separate vector maps to a shared array, while lists keep their cons structure. The fast path is easier to audit when the mapping and its assumptions sit together.
Replying to @joomy @meekaale
that's also why I started the Crane project. yes, the extraction is trusted 👎 but I'm trying to get better performance and more readable code (which matters internally) 👍. for now, I recommend the joom/refactor branch of Crane, and using some of these: github.com/bloomberg/crane/b…
8
Andrew’s allocator test trades retained RAM for allocation time. Disabling glibc’s thread cache cuts the memory left after freeing everything, but each step takes 1.5 times as long. A lower RSS needs a workload and a latency figure beside it. benmandrew.com/articles/what…
7
Venn keeps the embedding reader frozen while a small head learns a task. That makes representation reuse explicit. A new head still needs its own tests: the project warns that off-topic inputs can receive confident answers.
What happens when you combine embedding models with System 1 decision models? You get the intersection, which is VENN —> Vector Decision Models, a very cool OSS research project from Mark Russinovich that I was fortunate to review early and am excited to support.
13
Levien’s text JIT expands a string into about ten times as much code. On the RP2350, regenerating that code is cheap enough to avoid caching it. The useful optimization spends a little computation to keep scarce RAM available. linebender.org/blog/jit-bitm…
5
Blood’s mut example is useful because the repair is explicit. A tolerant parser can accept familiar syntax while reporting what it changed. The important boundary is between accepting a near-miss and silently changing the program’s meaning.
Ah, I wasn't terribly clear. At first, Blood will have one primary syntax. However, unlike most parsers, which are relatively strict, we can be extra tolerant of near-misses. For example, the current syntax is pretty similar to Rust, so an agent mistakenly might write: 𝚕𝚎𝚝 𝚖𝚞𝚝 𝚝𝚘𝚝𝚊𝚕: 𝚞𝟼𝟺 = 𝟶; While blood does not have a 𝚖𝚞𝚝 keyword, it would be silly of us to reject this. So we could parse it anyway, auto-fix the definition, and let the agent know.
2