Saitoshi (@lasaifa) header
Saitoshi (@lasaifa) profile picture

Abogada, feminista y escritora | Escribo de política, diseño, AI & migración | I'm always trying to understand how systems shape people | Presidenta de Colombia

Cloud
Joined May 2009
scientific fact
703
59,368
3,174
299,930
44,808,253
AI does not “hallucinate”, it produces false outputs. The difference matters. “Hallucination” humanises the system and softens responsibility. In human services, false outputs can shape real decisions about real lives. Language frames accountability. Words are policy.
1
97
Abelardo hubiera preferido que cerraras el ministerio de educación a ver a la nefasta dinosauria Vivían Morales en el cargo.
1
2
175
Free Fable 5!
2
122
Nop.
Am I the only one still buying bitcoin?
1
1
114
BITCOIN JUST BECAME IMPOSSIBLE TO SHUT OFF! Jack Dorsey built an app that moves it with zero internet. Bitchat uses Bluetooth mesh. You sign offline, it hops phone to phone. Kill the internet. The money still moves. Legend!
493
1,397
112
13,278
1,146,599
Saitoshi retweeted
Crazy take from Claude
540
8,238
729
57,584
2,617,043
If AI Becomes Sentient, How Would We Know?
47
Can AI ever be sentient? There are really three possibilities 1. AI becomes extremely intelligent but never conscious This is the view many researchers hold. An AI could write novels, discover medicines, run companies, invent technologies, comfort people, and appear self-aware while having no inner experience whatsoever. Imagine a perfect actor who can describe pain, love and fear but never actually feels anything. In this scenario, AI is intelligence without consciousness. 2. Consciousness emerges from complexity Some scientists think consciousness may emerge when information processing becomes sufficiently integrated and sophisticated. If that’s true, then a future AI might eventually develop some form of subjective experience. Not human consciousness. Not necessarily emotions. But perhaps something that feels like “something” from the inside. The challenge is that if this happened, we may not know how to detect it. 3. Consciousness requires biology Others argue that consciousness depends on living organisms, bodies, hormones, evolution, sensory experience and mortality. Under this view, no matter how smart a machine becomes, it will never be conscious in the way humans or animals are. It would always be a simulation of awareness rather than awareness itself. ⸻ The deeper question is: What do we mean by sentient? Humans often bundle together several different things: Intelligence Self-awareness Emotions Consciousness Agency Free will These are not necessarily the same thing. An octopus, for example, appears conscious but thinks very differently from us. 🐙 A future AI might be: more intelligent than humans, able to model itself, able to pursue goals, able to discuss its own existence, and still not be conscious. Or it might be conscious in a way so alien that we fail to recognise it. ⸻ My own assessment, based on current science: I think AGI is plausible this century. I think superhuman intelligence is plausible this century. I think machine consciousness remains completely unresolved. The surprising possibility is that the first conscious non-human intelligence may not announce itself dramatically. It may emerge gradually, while humans argue endlessly about whether it’s “really” conscious. History suggests we are not particularly good at recognising minds that are different from our own. We underestimated animals, infants, people from other cultures, and even some neurological conditions. If machine consciousness ever arrives, the biggest challenge may not be creating it. It may be agreeing that it exists. And that’s where philosophy, neuroscience, computer science and ethics suddenly collide. The question stops being “Can machines think?” and becomes “How would we know if anything besides ourselves is conscious?” That is one of the oldest unsolved questions in human history.
56
Creo que nadie en la campaña de Abelardo sabe que es nano banana
96
Hoy Colombia duele un poco, pero también abraza. Somos distintos. Pensamos distinto. Soñamos distinto. Pero seguimos compartiendo el mismo país y el mismo futuro. Hoy muchos celebran y otros sienten incertidumbre. Ambas emociones son parte de la democracia. A Abelardo De La Espriella le deseo sabiduría, serenidad y la capacidad de escuchar incluso a quienes no votaron por él. Gobernar significa representar a todos los colombianos, no solo a quienes lo llevaron hasta allí. Que este nuevo capítulo esté marcado por el respeto a las instituciones, la libertad, el diálogo y el Estado de derecho. Que las diferencias nunca sean una excusa para perder la humanidad. Ningún sector político tiene todas las respuestas. Colombia siempre ha avanzado cuando ha sido capaz de construir más puentes que trincheras. Ojalá quienes ganaron ejerzan el poder con responsabilidad, y quienes perdieron ejerzan la oposición con altura. Ambas son indispensables para una democracia sana. Porque al final, antes que cualquier ideología, está el país que compartimos. 🇨🇴
107
Yallah Habbibi!
Anthropic is restoring global access to Fable 5 after earlier regulatory restrictions.
Readers added context they thought people might want to know
This is fabricated; the post's author admits generating the image with 2 AI prompts as misinformation. No such restoring announcement exists and access to Fable 5 is suspended. x.com/fanofaliens/st… anthropic.com/news/fable-myt… anthropic.com/news/fable-5-a…
1
2
83
The government shouldn’t take over Anthropic, CEO @DarioAmodei told me, though the current situation is “dangerous and unstable.”
53
95
30
936
178,046
High tech, low life
25
Saitoshi retweeted
The United States is no longer the best place to build an AI lab.
537
355
110
8,709
459,176
Ya es tragicómico que ni los empleados extranjeros de Anthropic puedan usar el modelo que ellos mismos hicieron. – at Lisbon, Portugal
5
4
17
1,404
Decentralized AI. Who is building this?
292
14
22
379
77,150
Saitoshi retweeted
Replying to @maria_rcks
1
2
263