1,598 mentions · 615 chunks · 117 episodes
...have been too overwhelming so I decided not to. OpenClaw shows the raw power of LLMs when unleashed on your data. It's also self evidently, absurdly, catastrophically insecure. A friend described it this way: "It's basically a thought...
...r person noting that sandboxes are only a small part of the problem of securing LLMs. "I think most people focusing on securing these are focusing on isolation, but that's really step 0 of a step 3 process they'll come to understand a...
Clawdbot makes the danger of LLMs more obvious. In the past, "prompt injection" was hard to get even developers to think about. "That sounds like SQL injection, that thing we've solve...
Don't use LLMs to do things you could have done before, but faster. Use them to take on meaningful projects that you never would have attempted before.
Peter Wang calls LLMs "essence extractors." Notably, this is not just photocopying ideas. It requires judgment, nuance, and produces something structurally valuable and di...
...es Latent Space Engineering.[ci] It's why being polite and encouraging can help LLMs do better. You end up in different latent space basins by using the right words. LLMs do better when they think they can do it and are given positive...
Stuxnet was extremely expensive to create. But now LLMs have the potential to find the next Stuxnet for many orders of magnitude cheaper. Imagine a world where everyone could make their own Stuxnet to sic ...
...re was another step of the engineer actually implementing it. Now from the spec LLMs can just build it. That process of distillation of user need into software spec is more important than before, not less. Imagine if everyone had a pe...
...it required infinite patience to do it as a consumer, it wasn't viable. But now LLMs have infinite patience and you can deploy them to achieve your interests.
Chatbots are perhaps 1/10 or 1/100 of the actual value extractable from LLMs. What we see for chatbot subscription revenues reflects that lower efficiency of value. Already people who are using LLM coding agents are willing to...
...iew the output, and just share it with the world. When you have a dialogue with LLMs you can make something better than you could have made alone.
Centralized, singular LLMs must have a kind of bland beige aesthetic. Inoffensive to everyone and yet loved by no one. An internal consensus / average. When your work is edited...
...s-editor writing modality. If everything is written by a handful of centralized LLMs, we'll lose variety in our arguments in society.
When you discover writing is written by LLMs it feels like a betrayal. "Oh, I kind of like this. … wait, this was written by an LLM?!" If feels like "you tricked me" and you're embarrassed you f...
...llows them to run the world forever. There are a lot of companies in the age of LLMs that are positioned to have the "one rug pull to end all competition." Not saying any of them would do that… but they could!
What percent of a program's control flow is LLMs (vs normal code)? Agent startups assume it's 70%. If you took out the LLM the software wouldn't even exist. Another approach: assume it's 0-50%. That...
LLMs are a commodity, and if you act like that, a lot of things become more clear. The big model labs don't want that to be the case, but it's obviously t...
At some point in the quality curve of LLMs, "this sounds like an LLM wrote it" will be a compliment.[ct][cu][cv]
...anal engineering." We'll all just assume that of course most code is written by LLMs. Just like we assume all code today is compiled.
China is treating LLMs as a commodity, but the US isn't. The US is treating them like highly specialized IP. The Chinese approach is "AI is totally a commodity, we'll just ...