cedws

This individual project may die, but the AI rewrite movement is only accelerating, and for many people there’s a feeling of vengeance against the software companies that pushed AI so hard and treated their workers as disposable. That feeling is incredibly potent for motivation.

Anthropic just blocked REA but it doesn’t matter, we’ll just lean harder into the Chinese models.

Ironically your comment demonstrates exactly what I just said. Political polarisation plays into the hands of the elite, you're just helping to fuel it with this kind of tone.

It will be hard to keep everyone on the same page because if this gains any traction social media will ratchet up the propaganda and play the divide and conquer strategy. There'll be accusations of people joining movements like this of being communists and that will kick it off. Same strategy being used across the board.

IIRC Claude has some kind of mechanism where even though the system prompt is sent from the client, if it’s modified it can trip tamper checks and result in a ban.

The threat of damaging a nuclear reactor is grey zone, it’s unlikely to elicit the same response as a nuclear ICBM detonation.

Why is war never a factor in the nuclear power debate? Like nuclear is great, until you have missiles dropping on your territory. You can ask Ukraine about that. Zaporizhia has been held hostage since the early days of the war. Should the Russians have a manic moment they can make a Chernobyl volume 2 with conventional explosives.

Nuclear is a great peacetime technology, Europe is no longer in peacetime.

Fuck Anthropic, protecting corporate interests after themselves stealing from millions of people.

>generative SI

Eyeroll. Terrible that NASA must kowtow to the orange moron.

How does it resolve the prompt injection?

Same. Proton customer of about 8 years, I’m considering leaving, I don’t want to support such attitude. It’s not about the AI stance, it’s about attitude towards employees. Proton board, please fire him.

Random thought: when NK collapses, I wonder if there will be a rush to seize the crypto wallets. I guess the CIA will be there first.

For me connect time is not really a concern, it happens once when I open the app. In my project I'm wrapping agents using ACP and presenting an actual GUI, so latency doesn't really have a perceptible impact. Running Claude Code in a console over any kind of remote connection is gonna feel bad.

Rather than for network performance I'm using iroh because I don't want to need a direct network path to the machines running my agents. Yeah I have Tailscale but that's more work to set up and I don't want my app to depend on users have a TS network.

I've been building something similar except it uses iroh for P2P instead of SSH. Would really like to use this but I don't want to use SSH.

edit: nevermind I see it presents a console rather than wrapping the agent with a native GUI. This won't work well on high latency connections.

Trains shouldn’t even be manually driven in 2026. We have self driving cars now which is a way harder problem to solve. Relying on a system of signals and driver attentiveness is archaic.

Someone who isn’t me tried using GLM to crack some software, it wouldn’t comply. They had to use the abliterated version. I wonder how much of the Chinese models’ safeguards are a result of distillation, versus intentionally implemented safeguards.

I’m itching for more smart cheap models. I have a mid level Claude and ChatGPT subscription and I’m blasting through my usage, and I’m too cheap to pay for the higher tier.

I use DeepSeek too but it’s more of a workhorse model, requires a lot more guiding and sometimes struggles to get where I want it.

There was a time when women had more time to recover, time to look after the children, and they decided they didn’t want that and instead wanted to go into careers alongside men.

Of course, women should be able to do so if they choose. But the privileges that working households lost by both parents being in the workforce were never compensated for in another way. It expanded the labour pool, much to the delight of corporations that could push down wages. We went from one income being ample to support a family to two barely being enough. Now, as you say, the average family has the privilege of paying mountains of money to someone else to raise their children. Women became independent from their husbands but dependent on their bosses instead.

A linter is a program, a compiler is a program.

I don’t get why Anthropic hasn’t already vibe ported all of this critical C code to Rust. If they claim LLMs can autonomously do ‘good’ language ports like they did with Bun, there’s pretty much nothing better they could do for the world’s security posture than eliminate dangerous C code. Why even have these cyber programs and then not do the most obvious beneficial thing possible?

You’ll lose your job, but you can generate slop images. Have you said thank you yet?

Then they ushered in an age of garbage by removing the scale altogether.

The apps I've been interested in have attestation and reverse engineering countermeasures that block Frida.

Why not use agent forwarding?

Most interesting stuff is pinned unfortunately.

Sounds like it’s time for some involuntary open sourcing.

You aren't supposed to be moving SSH keys. They should stay on the machine they were generated on. The less accessible they are the better.

I really wouldn't rely on an SSH key as the sole way of accessing a machine.

Yes, and if Anthropic are trying what people are saying they are, that would be intentional. $1B could be reached quickly if cyberattacks supported by AI count.

Could you please share your methodology?

> “There were a number of posts that targeted me...It was unpleasant and I was targeted by a child I had no way of dealing with.”

Oh my god, how pathetic can you be. Seriously?

Also the part where you have three meals a day prepared for you, children’s needs taken care of, laundry washed ironed and folded, house cleaned. Working “hard” as a billionaire doesn’t mean shit.