tenso

Says only “decels” would tell you to pause the exponential AI progress we are witnessing now. A slogan-style stance against pausing, with no argument about specific risks. Full post text inspected via X’s public embed endpoint.

Quoting a post about frontier-lab employees signing a letter that asks the US government to support an international effort to slow frontier development, the account asks you to imagine trying to slow down a future where basic human suffering would not exist. The letter and the quoted post belong to others. Full post text inspected via X’s public embed endpoint.

Lists slogans it attributes to slowdown advocates (“data centers are bad”, “accelerating AGI is bad”, “pace the frontier”, “gatekeep access to progress”, “i wish i had a button to slow down everything”) and says these people will go down in history as the ones who tried to destroy the world. Hyperbolic rhetoric against slowdown advocates, not an analysis of their arguments. Full post text inspected via X’s public embed endpoint.

A one-line post saying AGI will not kill us but the politics around it will. Locates the danger in human politics rather than the technology; it does not say which political outcomes it means. Full post text inspected via X’s public embed endpoint.

Says the goal now should be to build weird new forms of computers, that abundance and AGI are already here, that you can hardwire an LLM into silicon, and asks whether Steve Jobs would build yet another AI agent that answers in Slack: “think outside the tokens”. A builder’s provocation in rhetorical form, not a measured capability claim. Full post text inspected via X’s public embed endpoint.

Says humans are now the minority on the internet, that the goal of AI was to create more humans, not more machines, and that “we are doing this wrong”. A concern about direction; it does not say what should change. Full post text inspected via X’s public embed endpoint.

A two-sentence post committing to open source and saying nobody can pace it, a jab at “pace the frontier” proposals. A stance, not an argument about misuse. Full post text inspected via X’s public embed endpoint.

Says aligning humans to humans is a hundred times harder than aligning AI to humans. A quip about where the hard problem sits, not a claim that AI alignment is solved. Full post text inspected via X’s public embed endpoint.

Says that if frontier labs really want to save humanity and make sure AI does not harm us, they should make all their model weights open, and that there is no other way. A demand for open weights framed as the safety answer; it does not address misuse arguments. Full post text inspected via X’s public embed endpoint.

Riffing on a quoted post (“if you understood how llms worked you’d know we don’t have true ai”), the account says that if you understood how LLMs worked you would know AI won’t kill us, only humans would. The quoted line belongs to @137beans. Full post text inspected via X’s public embed endpoint.

Says a robot has not even started doing its laundry, yet people want to spread “AI is going to kill us” psyops. Mocks AI-risk messaging by pointing to immature robotics; not an engagement with specific safety arguments. Full post text inspected via X’s public embed endpoint.

A list of ambitions: life has to become multi-planetary, data centers need to go to space, we have to terraform Mars, contact aliens, achieve true telepathy and own private spaceships, ending with “humanity has to accelerate collectively”. A techno-optimist vision statement without timelines. Full post text inspected via X’s public embed endpoint.