A Z Mackay (@azmackay) header
A Z Mackay (@azmackay) profile picture

Writer. Speculative fiction (Future Tense) and reported nonfiction on AI and safety-critical systems. Writing Safe Enough as Adam Mackay: https://nitter.cf/t.co/uQxcKroBrF

London, UK
Joined December 2008
Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)Photo by A Z Mackay (@azmackay)
Fiction and independent essays: azmackay.com I also write as Adam Mackay. Co-author of Embedded Software Testing; researching Safe Enough. Reporting, collaborative nonfiction and ghostwriting: adammackay.com/writing
1
292
ALTERNATE FRIDAY #003: The Moon, 1994. In a parallel universe, human lunar exploration didn't end with Apollo 17 in December 1972. By 1979 there was a permanent base at Tranquility, by the late 80s Britain had sent its first engineers (and you could post a card home from the Moon, second class, six to eight weeks), and by 1994 the Moon was where we built the ships that would go further. The first of them left lunar orbit that summer, and the captain's order on departure was just three words long... "Make it so." --- Welcome to 'Alternate Friday'. Follow me for a new alternative technological history every Friday ;-) #AlternateFriday #AlternateHistory #Apollo #SpaceExploration #RIPPicard
1
122
ALTERNATE FRIDAY #002: Artificial Intelligence, 1865. In a parallel universe, AI agents arrived powered by steam. The British Parliament responded with the Automatic Agents Act. Commonly known as “The Red Tie Act”. Every machine required continuous human supervision. Maximum permitted output: 212 words per minute. Every supervisor required a red tie. The future had arrived. Suitably dressed for accountability. Who decides what’s safe enough—and how fast it can go? #AlternateFriday #AI #AlternateHistory
2
247
For Safe Enough, I am exploring how AI evaluation evidence informs deployment decisions. Which public case studies show the reasoning and uncertainties? Project and contact: adammackay.com/writing
1
149
In 2025, experienced developers using AI were 19% slower at real tasks. But they believed they were 20% faster. That perception gap is the entire hype cycle in miniature. (article link in bio). #AI #HypeCycle #TechPredictions #FutureTense #CollectiveDelusion
2
3
425
Cambridge Dictionary's word of the year is "parasocial." 72% of American teens have used AI companions. A third find them as satisfying as real friendships. I wrote about what that reveals. (link in bio)
1
144
We all see the manipulation now. We correct less. We rage less. We just… scroll, exhausted. Meet the dropouts who quit and paid with their social lives (tracked in tragic spreadsheets). Meet the optimizers who hate the game but master it anyway (irony audits included). The machine doesn’t want you ignorant. It wants you exactly this aware. ft.azmackay.com/p/when-knowi…
1
121
A Z Mackay retweeted
" What’s harder to measure is whether anyone is thinking, or if we’ve simply built a society that no longer requires it for most roles. "
With Gemini 3's release ushering in even more powerful AI tools, the societal "sorting" between independent thinkers and AI-dependent performers is accelerating. It's reshaping education, jobs, and class divides. ft.azmackay.com/p/the-sortin… #AI #Gemini3
1
8
11
701
A Z Mackay retweeted
With Gemini 3's release ushering in even more powerful AI tools, the societal "sorting" between independent thinkers and AI-dependent performers is accelerating. It's reshaping education, jobs, and class divides. ft.azmackay.com/p/the-sortin… #AI #Gemini3
2
1
4
1,054
With Gemini 3's release ushering in even more powerful AI tools, the societal "sorting" between independent thinkers and AI-dependent performers is accelerating. It's reshaping education, jobs, and class divides. ft.azmackay.com/p/the-sortin… #AI #Gemini3
2
1
4
1,054
October 2025. A Harvard professor calculated the odds at one in ten quadrillion. NASA's scientists said "it looks like a comet." The gap between those two positions reveals more about human psychology than cometary composition.
1
1
125
Social pain uses the same brain regions as physical pain. Cosmic loneliness isn't metaphor. It hurts. And we're designed for a world different from the one we inhabit: Paleolithic brains confronting potential cosmic emptiness.
1
2
99
The rocks keep coming. Always just rocks. We keep looking, because accepting the silence means confronting what we're wired to deny. Read the full essay: ft.azmackay.com/p/the-third-…
1
93
Every major AI lab just pivoted to "reasoning models" that think for minutes before answering... except the thinking happens where you can't see it. Clean outputs, hidden deliberation, premium pricing for cognition you can't audit. When did we decide inscrutability was sophistication? ft.azmackay.com/p/the-tortoi… #AI #ArtificialIntelligence #TechCriticism #OpenAI #Anthropic #ReasoningModels #FutureTense
1
209
Thinking is becoming expensive. Performance is becoming free. The two-tier system is being built in real time through "personalized learning." You have 18 months before this becomes obvious to everyone. ft.azmackay.com/p/the-sortin… #futuretense
2
142
TICKET #847392 STATUS: PENDING ISSUE: National financial crisis WAIT TIME: 6-8 months "Have you tried restarting your economy?" Welcome to sovereignty as a service. ft.azmackay.com/p/geopolitic…
1
128
You're Not Addicted, You're Subscribed ft.azmackay.com/p/youre-not-…
1
2
647