There once was a time, over 30 years ago, that I was taught as part of my embedded software engineering training how to write pseudocode. It would appear that this technique is sheer perfection when writing code with Claude Sonnet AI. I feel like I have a superpower. CEO as SE x10 right now.

Interesting impact of the (assumed) roll out of Microsoft Defender Safe Links with new sandbox VMs for link scanning, as customers with big MailChimp subscriber lists can DDoS themselves with those VMs. Causes chaos with MailChimp click through stats as well. MailChimp not filtering the VMs IPs.

Interesting job at the National Gallery, London, 'Data Specialist - Heritage Science Digital Services Project' https://jobs.theguardian.com/job/9451812/data-specialist-heritage-science-digital-services-project/ £40k #MuseTech

Interesting read : lil.law.harvard.edu/century-scal... especially the Amazon note; I often think that Cloud based Digital Preservation is relying on companies who by their own admission might only last, what, 30 years? Should UK heritage preserve to offline, unpowered storage as the alternative?

Exquisite exhibit note writing here

Paid our ICO annual registration fee. Every year makes me wonder why, the Government's Information Commissioner's Office themselves don't use GOV.UK Pay but instead use a third party gateway. You'd think, for DPA , they would be *the* use case for Gov Pay ?

AR musings. Over a decade ago we user Layar , mesh based computer vision and 2D museum object recognition to perform collection database lookups. Can the same, for 3D objects, training e.g. Llama vision LLM and a non-native (i.e. web based / html) approach be taken? Seems still a long way off.

My first post shall be me marvelling at the discovery that each username on here is a subdomain of the federating server, so I am orangeleaf.bsky.social , which is rather jolly.