One of the best things about having a 3D printer on hand is being able to print off parts like this little Noctua A4x10 fan mount for my 430-16i HBA. Only took about 20 min.

#Homelab #3DPrint

If CEO swaps say anything, switching to the CEO known for Acquisition is definitely a sign. It seems my favorite password manager Bitwarden might not be long for this FOSS world, and they removed the " always free " wording on their site.

https://www.fastcompany.com/91542655/bitwarden-scrubs-always-free-and-inclusion-values-from-its-website-as-longtime-execs-step-down

Looks like Hacker News was already on top of this https://news.ycombinator.com/item?id=48147637

#privacy #security #passwords

Wonderful news for anyone who's talking to people across ecosystems, which is probably everyone. Apple and Google finally got to working together to roll out end to end encrypted RCS which means hopefully RCS will continue to replace plain text unencrypted SMS and maybe a bit more quickly. This is a huge improvement.

#privacymatters #wins

https://www.apple.com/newsroom/2026/05/end-to-end-encrypted-rcs-messaging-begins-rolling-out-today-in-beta/

Modern cars are basically rolling computers that send back significant telemetry, and that generally places them under privacy regulations. GM just settled and agreed to pay California $12.75 million after investigators found it was allegedly collecting and selling detailed driver data from OnStar users without proper consent.

#privacymatters

https://calmatters.org/economy/technology/2026/05/gm-record-california-penalty-onstar-data/

AI is a tool, but you probably only need a small hammer not a sledgehammer to hang a picture. Whether you're paying for tokens, or waiting for your 40-60 token/s GPU in the lab to process, that's when you may have realized what some corporate budgets are coming to realize, token spend is expensive and I should probably have my model write more ActivePieces/cron/ n8n automations so we're not wasting the compute.

https://www.benzinga.com/markets/tech/26/04/51828848/ubers-anthropic-ai-push-hits-wall-cto-says-budget-struggles-despite-spend.

#ai #llm #uber

Earlier I posted about how AI can't logic, and math is an example. Well clearly I was proven right today when booking hotels, from Perplexity: €1,157.56 (about $1,338-$1,355 USD) is not cheaper than $1,256.27 contrary to what the #Tokens believe.

#ai #perplexity #fail

A very real problem is relying on LLMs to do logic. LLMs fundamentally can't do logic, they are great tools for use cases where the underlying tokenization lends itself, like translation and code generation, but models on their own without tools still struggle with logic tasks like math. In the next years I can see audits being more focused on human in the loop review, especially for regulated areas that already generally require it.
#AI #LLM #security #privacy

https://arstechnica.com/ai/2026/04/research-finds-ai-users-scarily-willing-to-surrender-their-cognition-to-llms/