Your periodic reminder that #AI cannot even *see* the #future, much less *be* the future.

No matter how much #data a model trains on, it cannot create anything #original.

This is among the reasons why AI cannot replace #people. Every #job, even the simplest, requires the person doing it to be able to respond to something new and unforeseen.

This is something people often get wrong. We use the word #accident because sometimes things go wrong despite an individual's or group's #training and #experience. This use of the word "training" is different from what #techbros mean when they use it to describe their AI #models. Which is *another* reason why it can't reliably do what people do.

Původní slib zněl lákavě a jednoduše: přesuňte své IT do cloudu, ušetříte a zbavíte se starostí. Realita dnešního byznysu je ale mnohem složitější. Trh čelí bezprecedentnímu zdražování hardwaru, firmy střízliví z neřízených migrací a objevují se fenomény jako cloudová kocovina. Do toho všeho...

Tón: : mírně negativní
#česko #gdelt #cloud #data #uměláInteligence

https://archiv.hn.cz/c1-67879930-cloudova-kocovina-firmy-prepocitavaji-naklady-na-ai-cloud-i-digitalni-suverenitu

I wonder how big the largest connected component (LCC) of the Fediverse social graph is, where each node is a person and each edge is a follow relationship.

FediIndex shows 18,400+ user accounts. My guess is maybe 10% of them are unreachable through the follow graph.

It'd be interesting to find out!

#Fediverse #data #statistics #NetworkScience

https://fedi.wrm.sr/

The correct number of orgs that a service should share my data with is zero. Zero (0). None other orgs.

#privacy #data

Prosecutors in #Lithuania have opened a criminal investigation into a #data breach at the Centre of Registers (VIRC). More than 600,000 records may have been copied illegally from VIRC's #realestate and #business registers.

https://viabaltica.fi/lithuania-data-breach-at-state-registry/

I spent 50 hours drawing a line graph (dougmacdowell.com)

https://www.dougmacdowell.com/50-hours-to-draw-some-lines.html

#datascience #datavisualization #visualart #visualization #data

A rare network perspective on lack of access to data:

"Modeling the impact of research data unavailability on science."
https://www.sciencedirect.com/science/article/pii/S1751157726000465

"We analyzed how the loss of access to research resources propagates through interconnected bodies of scientific work. Our results show that information loss is not merely a local reproducibility issue but a structural phenomenon with system-wide implications…The unavailability of research data imposes a measurable efficiency loss on the scientific system…Preserving or restoring access to datasets associated with structurally central publications may prevent cascading losses that far exceed the local cost of data recovery. This insight provides a quantitative rationale for prioritizing preservation efforts based on network position and propagation risk."

#Data #OpenData #Preservation #ScholComm