ty13r

@TFTC21 @bitcoinplebdev How very Apple of them

Anthropic just killed cheap Claude access for OpenClaw and other third-party harnesses. Starting April 4, subscriptions no longer work outside their walled garden.

@shawmakesmagic assuming it’s not already 😅

Cryptography has a much scarier and shorter timeline than AI Likely that all encryption will be broken by state level actors before sovereign AGI

The speed at which science and research is advancing right now is the real issue. The real threat is when, it’s clearly on the horizon and another few leaps in advancement and we might be there. Private companies have much easier paths to upgrade their cryptography, bitcoin moving to PQ is going to be terribly messy.

@lopp the real question is wether or not ossification is going to be an issue when push comes to shove?

2 new quantum computing papers just dropped. Is crypto cooked? Google says they designed quantum circuits that could break ECC in a few minutes with 500,000 physical qubits: a 20-fold reduction from previous work. Oratomic says they could break ECC in a few days with 26,000 neutral-atom physical qubits. These papers both show advancements in algorithmic efficiency and quantum computing theory, but one should not overlook the assumptions underlying these claims. The authors have improved upon techniques shown to work at small scale, but we have no proof they can be scaled up. In other words, we're at the stage where scientists have created a few transistors but are still trying to figure out how to fabricate a fully functioning silicon chip with tons of transistors working together simultaneously. Progress is clearly continuing. How long do we have before a cryptographically relevant quantum computer can be built? That's still anyone's guess.

@harper very plausible

🤔

talking to claude code about the anthropic claude code source code leak:

@AlexFinn They’ve optimized their product development cycle around AI from the ground up. Their only bottleneck is figuring out what to build and when to build it.

The velocity in which Anthropic ships is unlike anything I've ever seen I'm sorry, but we have to assume they have access to an AGI like model that nobody else in the world has correct? They are simply holding it back so they can ship industry changing updates daily?

@RealJayWilliams All he had to do was hold the ball 🤦‍♂️

Watch tape of the last play in Duke vs UConn…. - play design? - baseline & sideline as extra defenders - Evans (best FT%) not in the frame - Cayden (turn, face, stay stationary & analyze - Karaban intelligent read - Evan’s hands low - Mullin’s dagger

I think @thsottiaux philosophy regarding the harness is the optimal path. “You’re scaffolding it in a way where you want to remove the scaffold over time because the model is able to stand on its own and it’s all about figuring out when is the right time to remove pieces of the scaffold.”

Harness engineering is as important as model capability scaling. AI Agents are 50% a harness story. "Natural-Language Agent Harnesses" proposes moving harness logic out of code and into the native language of LLMs: Natural Language. Turns the agent harness into a portable artifact that you can experiment with, improve and execute in a shared runtime. Agent harnesses will mature into a first class citizen of the AI Agent ecosystem. We will have people who are Agent Harness engineers and experts.

Now that I read so much content with em-dashes — I’ve started using them. But I don’t want my content to be considered AI slop. What to do? 🤔

@davemorin Is this how the frontier lab bubble pops?

Team Local Inference

@katibmoe I don’t understand the “open knowledge” marketing spin

Introducing One. The simplest way to connect and monitor AI agents to hundreds of apps. And we’re open-sourcing the world’s largest integration database powering it: 47,000 agentic actions across 250+ apps. RT + comment “One” for access & 1M free API requests/month.

@ziwenxu_ i was hoping we had plateau’s on spam calls and text messages but…

@_ty13r If you are my target audience.. I will spam you until you pay 🤣

Your OpenClaw agents can now text you AND make phone calls!! Most people don't realize what this actually unlocks. It means your agents can now scrape any site, anywhere, and pick up the phone to start selling your services or product. No human in the loop. Fully autonomous.

@boxmining A self-organizing distributed knowledge graph.

To all builders, What are you currently building with AI?

@shawmakesmagic Maybe this is what AGI was supposed to be?

The agent framework that wins will literally be a while loop, btw The existence of agent frameworks is existential proof that we don't have AGI yet, in the limit they are training wheels for stupid GIs

@mikehostetler Code is a great medium for doing things. It’s a terrible medium for knowing things.

Breaking: Meta researchers invented a perpetual motion machine with full energy conservation redefining the laws of thermodynamics Might as well pack it in, humans are cooked

Powerful local LLM models on phones in the years to come…

Wow

@Volland84 been following your writings around metagraph dbs and agent memory for months. I’d love to chat some time to share notes. Please feel free to DM me.

@masonic_tweets and that’s a good thing?

A 12% layoff is nothing for a large corporation. At any point in time you can reduce by 20-30% with no negative impact. AI should push this to 50-75% with net positive impact.