Miguel de Icaza ᯅ🍉

François Chollet, creator of Keras: “To be clear, at this time and for the foreseeable future, there does not exist any AI model or technique that could represent an extinction risk for humanity. Not even in nascent form, and not even if you extrapolate capabilities far into the future via scaling laws.

The Terminator scenario was science fiction back then (the movie was released almost 40 years ago!), it's science fiction today, and it will be science fiction in 40 years.”

Jun 2, 2023 at 12:23pmIvory for iOS
Show threadJanne MorenJun 2, 2023

@Migueldeicaza
"The Terminator scenario was science fiction back then (the movie was released almost 40 years ago!)"

No! It was just a few years ago! I'm still young, dammit!!

Show threadKhalid ⚡Jun 2, 2023
@Migueldeicaza True… but wait till they attach a machine gun to an algorithm. It doesn’t have to be smart to kill people.
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@khalidabuhakmeh you could attach a machine gun to excel today
Show threadKhalid ⚡Jun 2, 2023
@Migueldeicaza Excel is a weapon of a mass destruction
Show threadGaryJun 2, 2023
@khalidabuhakmeh @Migueldeicaza So that's why it likes to show #ERROR# so much with formulas. Trying to send the user over the edge.
Show threadJeff C. 🇺🇦🇬🇱Jun 2, 2023
@khalidabuhakmeh @Migueldeicaza A cold, calculating killer.
Show threadallisonJun 2, 2023
@Migueldeicaza @khalidabuhakmeh arguably "excel with machine gun attached" is just... capitalism
Show threadAlex T NewmanJun 2, 2023
@Migueldeicaza @khalidabuhakmeh we need to institute a pause in pivot tables. Just until we figure out what’s going on
Show threadKhalid ⚡Jun 2, 2023
@alextnewman @Migueldeicaza
Show threadMatt EricksonJun 2, 2023
@khalidabuhakmeh @alextnewman @Migueldeicaza relevant Micro SFF: https://mastodon.art/@MicroSFF/110474992323422395
Micro SF/F by O. Westin (@[email protected])

"We made a simulation," the colonel said. "Did you?" the robot said. "Well, a thought experiment." "And what did you think?" "That you could, in theory, kill someone." The robot looked at its bonds. "I was thinking the same thing." "That you could kill someone?" "That you could." #MicroFiction #SmallStories #TootFic

Mastodon.ART
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@matt @khalidabuhakmeh @alextnewman nice!
Show threadNick TuneJun 2, 2023
@khalidabuhakmeh @alextnewman @Migueldeicaza What was in Untitled 1?
Show threadKhalid ⚡Jun 2, 2023
@nick_tune @alextnewman @Migueldeicaza the meaning of life itself
Show threadNick TuneJun 2, 2023
@khalidabuhakmeh @alextnewman @Migueldeicaza ok, so nothing important.
Show threadLars Marowsky-Brée 😷Jun 2, 2023
@khalidabuhakmeh @alextnewman @Migueldeicaza Capitalism. This spreadsheet describes capitalism.
Show threadPēteris KrišjānisJun 3, 2023
@Migueldeicaza @khalidabuhakmeh or /dev/random
Show threadGabriel Adrian SamfiraJun 2, 2023

@Migueldeicaza and then stuff like this pops up:

https://nypost.com/2023/06/01/ai-enabled-drone-killed-human-operator-in-simulated-test/

AI-licensed drone killed human operator in surreal military simulated test

“We trained the system – ‘Hey don’t kill the operator – that’s bad,” Hamilton explained.

New York Post
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@gabriel fake: https://www.theguardian.com/us-news/2023/jun/01/us-military-drone-ai-killed-operator-simulated-test?CMP=share_btn_tw
Show threadGabriel Adrian SamfiraJun 2, 2023
@Migueldeicaza ok. Adding nypost.com to my AdGuard home.
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@gabriel sadly it got reported in other more reputable places as well
Show threadGabriel Adrian SamfiraJun 2, 2023

@Migueldeicaza showed up in my google news feed this morning.

I wish journalism went back to reporting well instead of reporting quickly...

Show threadJan KatinsJun 2, 2023
@Migueldeicaza interesting development: https://www.businessinsider.com/ai-powered-drone-tried-killing-its-operator-in-military-simulation-2023-6
Air Force official's story of killer AI was a hypothetical

The head of the US Air Force's AI Test and Operations warned that AI-enabled technology can behave in unpredictable and dangerous ways.

Insider
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@jankatins fake: https://www.theguardian.com/us-news/2023/jun/01/us-military-drone-ai-killed-operator-simulated-test?CMP=share_btn_tw
US air force denies running simulation in which AI drone ‘killed’ operator

Denial follows colonel saying drone used ‘highly unexpected strategies to achieve its goal’ in virtual test

The Guardian
Show threadAndrew ThekenJun 2, 2023
@Migueldeicaza "The Rise of the Machines" will be a lot dumber - faulty "Full Self-Driving" software that randomly misidentifies pedestrians as lane markers.
Show threadMarc KhouryJun 2, 2023
@Migueldeicaza I think the justified fear is not from the AIs themselves, but from the Moloch games that AIs can enable. You thought we couldn't deal with capitalism now? Wait til you see its next form
Show threadEnema CowboyJun 2, 2023
@Migueldeicaza Is AI an existential risk? Probably not. Can AI assist in creating dystopias? Most likely.
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@Enema_Cowboy well if that’s the baseline so can any SQL database
Show threadEnema CowboyJun 2, 2023
@Migueldeicaza The databases I work with would give Edgar Codd night terrors. 😆
Show threadMiguel de Icaza ᯅ🍉Jun 2, 2023
@Enema_Cowboy :-)
Show threadCF Bolz-TereickJun 2, 2023
@Migueldeicaza @Enema_Cowboy yup. I mean, punch cards managed
Show threadEnema CowboyJun 2, 2023
@cfbolz
IMDb: : Colossus: The Forbin Project
https://www.imdb.com/title/tt0064177/
@Migueldeicaza
Colossus: The Forbin Project (1970) ⭐ 7.1 | Sci-Fi, Thriller

1h 40m | M

IMDb
Show threadStefano ZacchiroliJun 2, 2023
@Migueldeicaza indeed the most scary part here is being remembered that the Terminator movie is 40y old. Please don't so it again!
Show threadadamcrussellJun 2, 2023
@Migueldeicaza this is the most common opinion of anyone who has done actual hands on work in AI.
Show threadAppApeJun 2, 2023
@Migueldeicaza @thomasfuchs “extinction risk” is doing a lot of heavy lifting here. The terminators didn’t pose an “extinction risk” either - humanity found a way to fight on. I also need to hear a LOT more about his reasoning - not enough to just say it. At its core nature AI is things happening that we aren’t capable of being aware of (and therefore not in control of). Yes, many things are like this- but we didn’t consciously decide to create them.