Milos Borenovic

OpenAI on X: "Announcing The Stargate Project The Stargate Project is a new company which intends to invest $500 billion over the next four years building new AI infrastructure for OpenAI in the United States. We will begin deploying $100 billion immediately. This infrastructure will secure" / X Announcing The Stargate Project The Stargate Project is a new company which intends to invest $500 billion over the next four years building new AI infrastructure for OpenAI in the United States. We will begin deploying $100 billion immediately. This infrastructure will secure

Is it just me or this reads like Manhattan Project II?

x.com/openai/statu...

BTW, total cost of Manhattan Project was less then 2bn USD (less than 30bn USD in today's USD).

Seems like it's getting gravity much better now.

I think applying past experiences to extrapolate future when there is a non-trivial chance of a singular change is a dangerous thing to do.

Even if they're aligned, if they're smarter then us, why wouldn't you listen to them telling you what they should build for us? Smells like losing the agency?

Help, can't decide where or what to start thinking about this

Anyone here thinking about doing a #phd? We have a healthy number of PhD studentships available for 2025/6 entry at the School of Computer Science, University of St Andrews, Scotland. These include 3.5 years of fees and a stipend, and are open to students from all over the world. Details below:

You've been driving the imaginary speed, at two orthogonal directions at once.

Future of Life Institute (FLI) is looking for a Head of U.S. Policy, ideally in Washington D.C, to work on US AI policy. The application deadline is December 22. The salary for this role is 150k to 240k.

Apply here: jobs.lever.co/futureof-lif...

They would certainly be easier to find and cheaper to hire than elite AI researchers. 😂

But yeah, data labeling entering a renaissance with very smart people catering the training data.

That one I know, I'm gonna be a dog trainer.

AI Governance & Human Agency

As transfomative AI becomes more ubiquitous, it seems that diminishing and managing the 'great divide' will be a challenge on par with AI alignment.

If we don't get it right, we might end up as multiple races.

Sorry Darwin, Your Monkey's Going Bionic Bridging the Gap From Both Ends: A Conversation with Kristian Rönn

Natural evolution vs. silicon speed - how do we keep up? Kristian's answer might surprise you 🤖

agencymatters.substack.com/p/sorry-darw...

#AISafety #AGI #FutureOfHumanity #AIGovernance #HumanAgency #AgencMatters

His book "The Darwinian Trap" introduced me to reputational markets - but his staircase metaphor for human-AI coevolution really clicked with my thoughts on preserving human agency.

"Sorry Darwin, Your Monkey's Going Bionic" 🐒

Fascinating chat with Kristian Rönn about bridging the future human-AI intelligence gap.

This nicely encapsulates one of the most important yet easily missed properties of LLMs.

No explicit "HUMANS MUST REMAIN IN CHARGE" declaration, but these guardrails tell a story: They're thinking about how to keep meaningful human control as AI systems become more autonomous.

Baby steps, but in the right direction? 🤔

#HumanAgency #AgencyMatters

Looking for human agency in the EU's AI Code of Practice? 🔍

It's there, but you have to read between the lines:

- Human oversight requirements
- Independent expert reviews
- Whistleblower protections
- Risk-based controls

#AI #AIGovernance

What would be your typical use?

You can solve a part of the problem by using more grounded ones such as NotebookLM but then you'd need to look for your sources.

Another partial solution might be to encourage it to ask questions when unsure. This should reduce the hallucination rate.

Agency Matters | Milos Borenovic | Substack What if the greatest threat from AI isn't that it becomes too powerful, but that we become too passive? Click to read Agency Matters, a Substack publication. Launched 21 days ago.

Hello everybody! 👋

My first post here. After years of building AI systems that reached billions, I'm now writing about keeping humans in the game while we still have a say. Thoughts on human agency in the coming age of AI: agencymatters.substack.com

#AI #AIEthics #AgencyMatters