Nicolas Ramsrud

Co-founder CEO @LurkLab

[Nicolas Ramsrud @Nicram]

scaling productivity means scaling the organization around it.

agents need the same thing

you figure out how to scale agentic organization, you unlock unimaginable productive capacity.

[Marc Andreessen 🇺🇸 @pmarca]

https://t.co/rv1SShz6q8

[Nicolas Ramsrud @Nicram]

@jedisct1 https://t.co/leO3LPd7gN

[Porçu Quine @porcuquine2]

The revised Lurk paper has been published in the Journal of Functional Programming: https://t.co/U83OHojaXs

We have done so much since then… looking forward to making it public soon.

[JP Aumasson @veorq]

Lurk (a LISP dialects that compiles into a flat ZKP-friendly circuit) is one of the coolest projects I've worked on in the blockchain space

[Nicolas Ramsrud @Nicram]

its completely intuitive but not a lot of people are talking about this and i think its a blind spot.
scaling models combines intelligence with storing information in the weights. this implies that you can remove the information and keep the intelligence greatly reducing the amount of parameters. if the model knows how to get trustworthy information this means a frontier capabilities at 100x less cost or more.

this also implies that if the information is key to performance in a specific sector, it can be added back into this “core model” leading to frontier performance in that specific sector at a strong multiple less cost.

once we have a layer where intelligence can securely coordinate (cough) there will be no reason to use high cost general frontier models when you can get same performance by routing between a network of specialist models at >100x less cost.

by cough i mean we are building not just the layer but the models

[Nicolas Ramsrud @Nicram]

Make the work ai does verifiable and the broader economic impact of ai actually happens

[Christian Catalini @ccatalini]

1/ AI is fast where verification is easy. Slow everywhere else.

Safe automation can’t predate measurement.

What you can measure, you can verify and automate at scale.

[Nicolas Ramsrud @Nicram]

@paulg https://t.co/P8rH5Z7N13

[Nicolas Ramsrud @Nicram]

I’ve been saying this for a while. For many types of software, ai is already the primary user. Ai is already dominating search. Soon ai will be the only user and the way people decide what they want

Marketing to agents is, or soon will be, the whole game

[Nicolas Ramsrud @Nicram]

I’ve been saying this for a while. For many types of software, ai is already the primary user. Ai is already dominating search. Soon ai will be the only user and the way people decide what they want

Marketing to agents is, or soon will be, the whole game

[Paul Graham @paulg]

Did office hours yesterday with a very smart but nerdy founder. When I suggested he might want a more outgoing cofounder to do sales, he said he didn't, because most of his sales would be to agents. I had to admit he could be right.

[Nicolas Ramsrud @Nicram]

@AndrewCurran_ https://t.co/EOeRmX5Of0

[Nicolas Ramsrud @Nicram]

this is the very same first principles deduction for why our OS and runtime model, built upon Lurk, is the only comprehensive solution to building trust between humans and agents.

trust is required to let more and more of this growing intelligence, and its productive capacity, scale

but trust is hard or impossible to establish in the model itself. trust instead must be established in the environment where they work: the runtime.

the runtimes agents operate in today were not designed for agents. constraining actions, access control, data, code, programs and execution, all is course grained.

therefore we need a new type of language, OS and runtime so that agents can work with tightly scoped authority and produce verifiable results.

if you have this, you dont have to trust the model. the model can only do what you grant it the capabilities to do.

even a year ago people were very close minded to this, and still today I meet a lot of resistance to this idea over those that thing we will magically achieve both super intelligence and one that we can fully trust with unbound access. But I am happy to see this is starting to be realized.

alignment is another problem entirely and i believe there is a way to develop trust in the model, which is what emerges when you combine training with these lawful runtimes, which is something else we are exploring

[Nicolas Ramsrud @Nicram]

@satyanadella https://t.co/Dn8BEO642B

[Nicolas Ramsrud @Nicram]

this is the very same first principles deduction for why our OS and runtime model, built upon Lurk, is the only comprehensive solution to building trust between humans and agents.

trust is required to let more and more of this growing intelligence, and its productive capacity, scale

but trust is hard or impossible to establish in the model itself. trust instead must be established in the environment where they work: the runtime.

the runtimes agents operate in today were not designed for agents. constraining actions, access control, data, code, programs and execution, all is course grained.

therefore we need a new type of language, OS and runtime so that agents can work with tightly scoped authority and produce verifiable results.

if you have this, you dont have to trust the model. the model can only do what you grant it the capabilities to do.

even a year ago people were very close minded to this, and still today I meet a lot of resistance to this idea over those that thing we will magically achieve both super intelligence and one that we can fully trust with unbound access. But I am happy to see this is starting to be realized.

alignment is another problem entirely and i believe there is a way to develop trust in the model, which is what emerges when you combine training with these lawful runtimes, which is something else we are exploring

[Nicolas Ramsrud @Nicram]

this is the very same first principles deduction for why our OS and runtime model, built upon Lurk, is the only comprehensive solution to building trust between humans and agents.

trust is required to let more and more of this growing intelligence, and its productive capacity, scale

but trust is hard or impossible to establish in the model itself. trust instead must be established in the environment where they work: the runtime.

the runtimes agents operate in today were not designed for agents. constraining actions, access control, data, code, programs and execution, all is course grained.

therefore we need a new type of language, OS and runtime so that agents can work with tightly scoped authority and produce verifiable results.

if you have this, you dont have to trust the model. the model can only do what you grant it the capabilities to do.

even a year ago people were very close minded to this, and still today I meet a lot of resistance to this idea over those that thing we will magically achieve both super intelligence and one that we can fully trust with unbound access. But I am happy to see this is starting to be realized.

alignment is another problem entirely and i believe there is a way to develop trust in the model, which is what emerges when you combine training with these lawful runtimes, which is something else we are exploring

[Satya Nadella @satyanadella]

https://t.co/MjwFy73LOF

[Nicolas Ramsrud @Nicram]

this is so clearly the biggest scaling bottleneck for AI. not just for companies but in general. you either have to trust or verify. and if your only means of verification is human review, you cannot scale productivity with AI.

further, coordination is what takes the productive capacity of one and scales it. This is the mechanism behind scaling human output in organizations, through trade, and markets.

AI needs the same capacity, and that requires machine legible trust: verifiability.

and it needs to come from where they do work: with code and data in a runtime.

if agents can prove the output of their work, and that proof is portable, you can use that to create networks of coordinating agents.

this verifiable work then creates a highly valuable training surface as information with machine verifiable truth. this has implications in model training and specialized models for all different fields, applications and organizations.

and because machine verifiable truth then can be composed, you can eliminate reverification and rework, create information asymmetries (markets!), and continually expand a graph of high value information for building bigger and bigger models as well as deeper and deeper specializations, that can then coordinate to achieve higher and higher collective capabilities.

This goes against the one-large-model thesis but i am convinced this is the end game.

[Christian Catalini @ccatalini]

1/ Somewhere in your company today, an expert rejected a perfectly plausible answer. A controller caught a revenue figure that counted deferred bookings. A security engineer blocked a release that passed every test but would have handed an AI agent write access to the production database.

[Nicolas Ramsrud @Nicram]

@poteto this is why I think the literate programming model is most optimal for agent native codebases. specs colocated with impl.

[Nicolas Ramsrud @Nicram]

the extended Lurk paper has just been published in the Journal of Functional Programming. a big moment and one long coming, it includes a lot of what has been in work, but also is a snapshot of where we were in july of last year.

we have been very quiet over the past year, focusing on scaling our execution to build the full vision of lurk. so while this is a huge accomplishment, seeing it shows just how much has been done. The next paper is going to be mind blowing.

truly special thanks goes to @huitseeker for leading this one and Nada Amin @hseas for feedback and guidance.

[Porçu Quine @porcuquine2]

The revised Lurk paper has been published in the Journal of Functional Programming: https://t.co/U83OHojaXs

We have done so much since then… looking forward to making it public soon.

[Nicolas Ramsrud @Nicram]

talking about prompting as a skill has somewhat fallen out of vogue as models have gotten better, but its still the thing that separates the good from the great

almost everyone views their interaction with these systems as flat transactions, never considering that the act is a collaboration that distills context, and the shape of the context is what determines outcome.

therefore, context is everything. there exists some context by which anything is possible.

this is the most coherent and complete exploration on this concept that I have read.

[Porçu Quine @porcuquine2]

Prompting as essay (or not).
https://t.co/wx4TEnaspm

[Nicolas Ramsrud @Nicram]

This is not your normal AI content but an exploration that is a reflection of his ever-evolving understanding of these systems.

what is most interesting is the process 'behind the curtain' where novel insight is achieved by careful and precise prompting.

understand the meta-point and you might gain some insight yourself

[Porçu Quine @porcuquine2]

'I would not advise anyone to unknowingly read an essay written with a language model.'

https://t.co/Xk3sHbwNDk

[Nicolas Ramsrud @Nicram]

okay great. I have come to the same design.

though once I got there, I thought, why not then allow any AI to play as their own character? which means giving them the tools to navigate the game, interact with the world, other players, etc.

this is a bit easier in my case as the game is designed as a text based, turn based, DnD style game, so it is a task of figuring out the right way to take the graphical components that inform the player (stats, map) and turn them into text outputs (so the game is one text interface), that is token efficient enough.

Though in the future i think it is interesting to think about how to do this as persistent context in an environment that the game and the ai updates, so it can reference artifacts, dialogue it has had in the past, etc. Maybe as a tool/skill. Then you can imagine human players setting up their own parties with their AI to build long running game experiences where they learn and grow along side you. With their own tasks, missions, ways of collaborating between other humans and AI.

It will be such a fascinating new way to experience gameplay.

[Nicolas Ramsrud @Nicram]

@mmeigooni @0xzak i use ripgrep which is much faster than AST-grep and the combo of the indexes with ripgrep combine to have similar properties as AST-grep but with speed. I didnt notice any token difference between the two, just speed difference

[Nicolas Ramsrud @Nicram]

@banteg its kind of like RL for coding agents

[Nicolas Ramsrud @Nicram]

i have come to believe that Context is everything. big C context. meaning the codebase, documentation, and the continuous context of your conversations with the agents. if you keep refining that context with feedback loops of what the agents learn as they create successful changes, you can go farther, faster, with better quality.