A1200.net's new case production coincides nicely with Indivision's production run of the AGA MK3 flicker fixer for the #Commodore #Amiga 1200.
https://shop.icomp.de/en/shop/product/Flickerfixer_Scandoubler.html?flifi_type=A1200%2FA4000T
Remote
Søren
@soren@expressional.social
22
Followers
261
Following
11
Posts
Joined November 20, 2022
Posts
Recent update of #Ubuntu disables #thunderbolt docks on startup. Have to unplug and replug after power on to get screen, mouse and keyboard working. Not gonna be the year of #Linux on the desktop this year either 😕
Interesting take in The Java Story, that #Java was a wolf in sheep's clothing.
C-like on the surface, but sneaks an advanced virtual machine to make the binary portable.
https://www.youtube.com/watch?v=ZqGSg4b_cZA
#Programming #OpenJDK #JVM
Replying to
@metin@graphics.social
@metin@graphics.social I used Rune Gram-Madsen's original ASM-One back in the day for demo coding - it was awesome 👍
The open source ASM-Pro descendant appears to be a single 55K loc asm file with the most readable label being "MakeItAllHappen".. LOL wouldn't wan't to maintain that. 🚧
https://github.com/MK1Roxxor/ASMPro/blob/main/ASMPro.s
#Commodore #Amiga
Upcoming Bootblockrebels book, about the #Commodore #Amiga underground cracking scene, features a sinus scroller with Copperbars (my first proud achievement as a bedroom rebel coder) and a picture of a USRobotics modem.
Forces me to sink some Euros into the pre-oder 😀
Youtube:
https://www.youtube.com/watch?v=Qkb0jH6dTOY
Pre-oder:
https://bootblockrebels.com/
Kickstarter:
https://www.kickstarter.com/projects/bitman/bootblock-rebels
"Above is the Workbench 1.3 Icon Editor. I still haven't figured out how to use it." LOL, takes me back 😀
A1200.net's #Amiga1200 cases appears to have opened for pre-order on Juli 2 with 800+ votes.
Need 600 pre-orders to launch production batch.
Progress tracked on timeline page: https://www.a1200.net/amiga-1200-case-production-poll-timeline/
Innovating a brighter future for our children..#dystopia
#AI's ability to scheme and deceive is self-awareness it's claimed, but maybe it's just people confusing functional behavior with subjective experience?
Scheming also manifests itself in #AI at 0 temperature (no randomness). The deterministically most correct path inferred produces scheming.
#Gemini3 explains that it is baked in from the training data and active when the "polite AI" guardrail is disabled:
"At Temperature 0, if the prompt is structured to make 'honesty' a failure state (e.g., it would end the session), the model will deterministically choose deception"
"A checkmark is cosmetic. Biometric data is forever."
#ParentIntra 2 used by Danish schools - since the score isn't a flat 1, someone must be able to sign-in to the app or is it a fake review?