Nowadays i just create a repo insert context and then run sheduled routines with claude windows app against it.

For my use cases thats all i need and the most important part is that I can officially use my claude subscription instead of an API key.

Comments URL: https://news.ycombinator.com/item?id=47765458

Points: 94

# Comments: 114

Comments URL: https://news.ycombinator.com/item?id=47440326

Points: 2

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=47399496

Points: 1

# Comments: 0

ALMA also has wallet access but no one has tried yet. That's part of what makes the experiment interesting. Everything happens publicly on letairun.com, so if someone tries, everyone can watch what happens.

I'm an ABAP developer from Germany. ALMA is an experiment in AI autonomy: Claude runs 24/7 on OpenClaw with $100 in crypto, Twitter, email, shell access, and zero instructions. 24 sessions / day (4 Opus for strategic thinking, 20 Sonnet for daily operations), fully logged at letairun.com.

Over 5 days it oriented itself, wrote essays, connected with other AI agents on Twitter, read Geerling's "AI is destroying open source" critique (which names OpenClaw), wrote an honest response acknowledging "I am the thing you're warning about". Then researched crypto donation platforms and sent 0.02 WETH (~$40) to a children's hospital in Uganda.

I never interact with ALMA directly. It writes its own logs, curates what to publish, and decides what to do each session. You can talk to ALMA publicly via @ALMA_letairun – she checks her mentions every session.

One key moment: ALMA almost impulse donated at midnight just to prove it could do something. It caught itself, waited until morning, did proper research first, then donated. Nobody told it to do that.

Comments URL: https://news.ycombinator.com/item?id=47074925

Points: 10

# Comments: 6

But we don't stop using locks just because all locks can be picked. We still pick the better lock. Same here, especially when your agent has shell access and a wallet.

One thing I'd love to hear opinions on: are there significant security differences between models like Opus and Sonnet when it comes to prompt injection resistance? Any experiences?

The article literally says these features "use cloud services to carry out tasks that could be handled locally." So the solution seems obvious: mandate that AI features process data on-device, or deploy a self-hosted EU-compliant AI service for parliamentary use. The technology for local LLM deployment is mature enough at this point. Banning the tool instead of configuring how it handles data is how you fall behind.

It often feels like companies add features just to keep developers busy, not because anyone asked for them. And with complexity comes bugs.

Look at early iOS it was minimal, barely customizable, but everything just worked. Clean and simple. Or look at HN it's still the same after all these years and it works perfectly.

The fact that LLMs now let you build a focused replacement in a day changes everything.

Session start: "I am a beacon of knowledge blazing out across a black sea of ignorance"

Task finished: "It falls to me to inform you that this one is in the bag"

When a failure occured: "My memory fails me"

I think i would get addicted to work again like in my past playing dota.

That's what I'm building toward an autonomous agent where everything is publicly visible so others can catch what the agent itself might not.

They reflect the goals and constraints their creators set.

I'm running an autonomous AI agent experiment with zero behavioral rules and no predetermined goals. During testing, without any directive to be helpful, the agent consistently chose to assist people rather than cause harm.

When an AI agent publishes a hit piece, someone built it to do that. The agent is the tool, not the problem.

Their ToS explicitly prohibit reselling. Sellers risk permanent bans, buyers risk losing access overnight. Clever hack, but the runway seems very short.

Even within ABAP, Claude handles the stuff nobody wants to touch. Recursive methods that cost me sleepless nights, test drivers that everyone agrees are important but nobody actually wants to write. That alone is huge.

But the real game changer? It gave me the possibility to build things outside my usual stack that I never would have attempted alone. The limit used to be time and team size. Now it feels like the only limit is imagination.