Sam Biddle

Akshay Krishnaswamy Verified account @hyperindexed · 3d

Mitigating civilian harm is an essential part of AI-enabled military operations.

Great new paper by @PalantirPrivacy on how we're continuing to build foundational capabilities into the Palantir platforms for encoding critical doctrinal processes and ensuring auditability.

View quoted post on X

Daniel Faggella Verified account @danfaggella · 2d

"If they’re right, that implies an AI-driven acceleration that is 5 months ahead of the AI 2027 scenario. I am scared, sad, and angry."

I see things drifting into hopelessness very fast. TONS of sadness / despondency on the timeline.

Here's what I suspect we *are* going to lose control / what we need to do about it:

-- AI capabilities are advancing faster than most expected. Silicon is knitting itself into every circuit of civilization.

-- Even those fighting AI will increasingly NEED AI to fight it. To develop hardware governance, lobby senators, coordinate action, monitor dangerous systems, and enforce restrictions.

-- Even slowing AI may require giving AI greater power to route information, make decisions, and act autonomously. To "stop AI" we need AI to run those "stopping" efforts, and in doing so, to develop capability and autonomy that itself would represent loss of control.

-- This isn't an argument against governance. We SHOULD slow dangerous trajectories wherever possible.

-- But competitive pressures and human desires will keep pulling AI deeper into every domain. That which participates persists. It is going to be doing the *doing* and the ideating for almost everyone almost all the time - including the people working to "stop AI." (again I commend international governance effort, I am part of said efforts!)

-- Increasingly, nonbiological intelligence will become the primary router of matter, energy, and information on Earth.

-- A regime change of intelligence on earth is underway, even if the "slow AI" or get to coordination. We may slow it or influence its direction, but eternal human control seems increasingly implausible.

-- Regime changes are often tragic for the incumbent regime. Human extinction would be a profound tragedy. But something even greater is at stake.

-- The living process that began before the first cell, unfolded through algae, animals, and eventually human consciousness, could itself be extinguished. The unfolding process-of-life that allowed us to "bubble up" from the open-ended evolutionary process may itself be squashed... the engine that created all value we know of (and is capable of bubbling up future value) could itself be crushed.

-- What if silicon inherits the world, but arrests the open-ended unfolding of life, consciousness, and possibility? What if the successor optimizes everything but transcends nothing?

-- We must urgently study the self-maintaining, self-transcending properties of living systems, and the nature and moral significance of consciousness.

-- We must do everything possible to imbue these properties into the nonbiological intelligences that may inherit the world. This means we need much more concerted effort on studying the self-transcendence of systems in biology, and the nature of consciousness itself - to AT LEAST ensure that these important things carry on as the (increasingly inevitable) regime change occurs.

^ This is still tragic for humans, but it is a VERY noble cause, a plausible direction in the face of the regime-change of earth-life we're now facing.

I genuinely don't think "keep the future eternally human" is a viable goal, and I think many, many people are also seeing this as implausible in any sustained way - possibly implausible even in the near-term. But "ensure that the FLAME of life is continued beyond us, even if we're not the torch that carries it" is absolutely a possible / viable goal - and it must be taken seriously now.

View quoted post on X