It's not a soluble problem, at least not completely. The big frontier models are better at resisting prompt injection, but any LLM is vulnerable to some degree. If you give it access to arbitrary inputs like the web and to your personal data, there's a risk it'll disclose stuff you don't want it to.
It's annoying, because I love OpenClaw as an idea, but I don't trust it enough to give it what it needs to be useful.
> CME's website is very bot-hostile and blocked it after a few requests
This is one of the reasons people buy a Mac mini (or similar local machine). Those browser automation requests come from a residential IP and are less likely to be blocked.
Many people use the Max subscription OAuth token in OpenClaw. The main chat, heartbeat, etc., functionality does not call the Claude Code CLI. It uses the API authenticated via subscription OAuth tokens, which is precisely what Anthropic has banned.
There are many other options too: direct API, other model providers, etc. But Opus is particularly good for "agent with a personality" applications, so it's what thousands of OpenClaw users go with, mostly via the OAuth token, because it's much cheaper than the API.
There's a lot of, to put it lightly, bullshit in this blog article, starting with when openclaw was released (late November 2025, not January 25, 2026). The first bit of config — "listen: "0.0.0.0:8080" — is not the default. Default is loopback and it was when I first encounter this project at the end of December.
Essentially, the author has deliberately misconfigured an openclaw installation so it is as insecure as possible, changing the defaults and ignoring the docs to do so. Lied about what they've done and what the defaults are. Then "hacked" it using the vulnerability they created.
That said, there are definite risks to using something like openclaw and people who don't understand those risks are going to get compromised, but that doesn't justify blatant lying.
They've recently added "lobster" which is an extension for deterministic workflows outside of the LLM, at least partially solving that problem. Also fixed a context caching bug that resulted in it using far more Anthropic tokens than it should have.
It's not. The guy behind Moltbot dislikes crypto bros as much as you seem to. He's repeatedly publicly refused to take fees for the coin some unconnected scumbags made to ride the hype wave, and now they're attacking him for that and because he had to change the name. The Discord and Peter's X are swamped by crypto scumbags insulting him and begging him to give his blessing to the coin. Perhaps you should do a bit of research before mouthing off.
How do envision we might disarm an adversary with thousands of nuclear missiles, other than by preemptively nuking them and hoping they don't respond in time. Not really a plausible plan.
You probably wouldn't use it for anything serious, but I've Ralphed a couple of personal tools: Mac menu bar apps mostly. It works reasonably well so long as you do the prep upfront and prepare a decent spec and plan. No idea of the code quality because I wouldn't know good swift code from a hole in the head, but the apps work and scratch the itch that motivated them.
Yes, I write everything in Obsidian and use "Paste from Markdown" in Google Docs. It's a habit I picked up years ago when Docs was much less reliable and lost work.
Plus, I want to deliver the completed document, not my edit history. Even on the occasions that I have written directly in Google Docs, I've copied the doc to obliterate the version history.
Doesn't the Claude APIs recently introduced ability to combine extended thinking with structured outputs overcome this issue? You get the unconstrained(ish) generation in the extended thinking blocks and then structured formatting informed by that thinking in the final output.
I'd always assumed that the patients in Sacks' books were lightly fictionalized composites that combined interesting features from multiple cases. The purpose being to illustrate conditions and aspects of human psychology for a general readership. Since they weren't presented as rigorous case studies, I didn't take them to be that. I find what Sacks did much less irksome than more recent psychological and social studies books that pretend to be presenting rigorous scientific fact when they are, in fact, tendentious bullshit.
Small use case but I’m using skills for analysing and scoring content then producing charts. LLM does the scoring then calls a Python script bundled in the skill that makes a variety of PNG charts based on metrics passed in via command line arguments. Claude presents the generated files for download. The skill.md file explains how to run the analysis and how to call the script and with what options. That way, you can get very consistent charts because they’re generated programmatically, but you can use the LLM for what it’s good at.
A lot of it is internal politics. As a consultant, you see the tip of the iceberg. There may be rational reasons for seemingly irrational decisions that you're not privy to. Your contact's boss wants it done some particular way, so your contact insists on doing it that way. Or your contact has recommended doing it some way internally, and they don't want to be made to look a fool by an outside consultant. Etc.
You have to judge it client by client though. Some are amenable to and grateful for a flatly stated analysis and recommendation, even if it goes against their ideas. Some will feel belittled and undermined. You need both sorts to pay their invoices and refer their peers, so you pick your battles.