Alco Duck

@alcoduck.bsky.social

"...programs must be written for people to read, and only incidentally for machines to execute."

5/ In general, work that shortens feedback loops is the most valuable "metaproject" work that can be done on a project. Doesn't matter what feedback loop: testing, compiling, deploying, getting customer feedback, analyzing data, whatever

Vibe coding broke people's brains because they had a bad understanding of the software design process. The pop culture model goes something like this: start out with a sketch and then render that same idea in progressively finer detail. So when tools could "skip" to high detail, people went WOW.

It's interesting how "...or get left behind" never seems to be attached to a principle or practice in software development, only to technology. e.g., Teams who do big, infrequent releases really *are* getting left behind. But I don't recall any tech CEOs standing up on stage warning us about that.

From the same dev: “it’s also that I cannot see management being concerned solving the information visibility that keeps slowing us down. We’re rolling out AI coding tools for devs but no one thinks about the org silos that slow us down 10x more than any coding use case.”

Onlangs sprak ik weer een droeve ambtenaar die van alles zag verhuizen naar AI op Amerikaanse clouds. En daar kan je helemaal kapot aan gaan, want in je eentje hou je het niet tegen. Hier een hart onder de riem: het ligt niet aan jou, er gebeuren inderdaad absurde dingen: berthub.eu/articles/pos...

Ga niet kapot aan je eigen overheid - Bert Hubert

Onze overheid heeft regels opgesteld voor hun gebruik van de cloud, AI en andere algoritmes. En ze houden zich er bijna nooit aan. Dit wordt keer op keer vastgesteld, zoals door de Algemene Rekenkamer...

berthub.eu

Many EU member states are arguing for forcing WhatsApp to inspect all our photos w/AI. If the AI is in any “doubt” if it might be child pornography, your photo, location & other details get reported to Europol and a local police force. This is a terrible plan: berthub.eu/articles/pos... #chatcontrol

Chatcontrol 2025 edition in Brief - Bert Hubert

In short, led by Denmark, many EU member states are arguing for forcing WhatsApp/Signal/etc to inspect all our photos and links, using AI. If the AI is in any “doubt” if this might be child pornograph...

berthub.eu

It’s so frustrating that there is this illusion that we are all 49%/49% right now. Us versus them, good versus bad, and it’s all because of the BS that is the electoral system. It’s very likely 70% versus 30% and we’re just not seeing representative government.There’s just no common sense right now

Americans! Do not read this paper! * Contains the following forbidden words: Black, transverse, transition, diversity, inclusion * Posting scientist has pronouns in bio, and is probably vaccinated * Authors neglected to include five bullet points about what they accomplished this week

Astronomer Royal for Scotland | Catherine Heymans@astroroyalscot.bsky.social · last yr.

Congrats to @desisurvey.bsky.social - it's looking more & more like dark energy isn't Einstein's cosmological constant! If you've woken up to the news & are searching for the publications, they don't appear to be on arxiv 🤷‍♀️ but they are online. 🔭 #cosmology 🧪 ℹ️: data.desi.lbl.gov/public/paper...

A graph of the evolution of the dark energy (vertical axis) as a function of time (redshift on the horizontal axis) - showing an offset from the cosmological constant (a flat horizontal line).   Here is the figure caption from the paper:
FIG. 12. Comparison of the constraints on the equation of
state of dark energy using the CPL parametrization and a bin-
ning reconstruction approach, using DESI+CMB+DESY5.
The solid green line shows the best-fit w(z) based on w0 and
wa inference and the green contours around it represent the
68% and 95% confidence intervals. We also show in blue the
constraints from the binning approach [47], with the hori-
zontal bars indicating the bin width (which is fixed) and the
vertical bars representing the 1σ error. Additionally, we in-
clude in gray the 1D posterior for each binning parameter.
The ΛCDM limit corresponds to the horizontal gray dashed
line.