“OpenClaw is interesting, but will also drain your wallet if you aren't careful.
Last night around midnight I loaded my Anthropic
API account with $20, then went to bed.
When I woke up, my Anthropic balance was $O.
Opus was checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night."
Doing literally nothing, OpenClaw spent the entire balance. How?
The "Heartbeat" cron job, even though literally the only thing I had going was one silly reminder, ("remind me tomorrow to get milk")”
Continuation of twitter post
“1. Sent ~120,000 tokens of context to Opus 4.5
2. Opus read HEARTBEAT md, thought about reminders
3. Replied "HEARTBEAT_OK"
4. Cost: ~$0.75 per heartbeat (cache writes)
The damage:
Overnight = ~25+ heartbeats
25 × $0.75 = ~$18.75 just from heartbeats alone
Plus regular conversation = ~$20 total
The absurdity: Opus was essentially checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night."
The problem is:
Heartbeat uses Opus (most expensive model) for a trivial check
Sends the entire conversation context (~120k tokens) each time
Runs every 30 minutes regardless of whether anything needs checking
That's $750 a month if this runs, to occasionally remind me stuff? Yeah, no. Not great.”
Diligence is costly in executive attention, it is relatively rare that a major donor is using your acceptance of donations to get social cover for an island-based extortion operation
Either deliberately whitewashing the situation or completely missing the point of why people are mad at Epstein, Yud really can't help himself.
edit: Or depending on the timeline and the fact that 'prison time for soliciting a 14 year old' was on top of Epstein's wiki as early a 2016 he's explicitly saying they didn't mind that part with 300k on the line.
I’m planning on using this data to catalog “in the wild” instances of agents resisting shutdown, attempting to acquire resources, and avoiding oversight.
He'll probably do this by running an agent that uses a chatbot with the playwright mcp to occasionally scrape the site, then feed that to a second agent who'll filter the posts for suspect behavior, then to another agent to summarize and create a report, then another agent which decides if the report is worth it for him to read and message him through his socials. Maybe another agent with db access to log the flagged posts at some point.
All this will be worth it to no one except the bot vendors.
How did molt become a term of endearment for agents? I read in the pivot thread that clawdbot changed it's name to moltbot because anthropic got ornery.
It's so blindingly obvious that it's become obscure again so it bears pointing out, someone really went ahead and named a tech company after a fantasy torment nexus and people thought it wouldn't be sketch.
Somehow I doubt sneerclub turning out to be gigachad central is going to do wonders for curing their persecution complex, but I'm here for it, I even have a couple of (really local) combat sports tournament medals to show for from back in the day.
Self-hating nerd may sound like a quip, but he is pretty specifically painting us as tribe traitors.
Remember, for them, the male nerd is a vulnerable minority, and that they haven't been granted protected status yet is possibly the greatest injustice of our time.
We're basically supposed to be fucking up their chances for finally instituting a society-wide word-count based sex redistribution scheme by cozying up to the Man normies who think cults are bad and don't appreciate race science.
woo takes about quantum mechanics and the power of self-affirmation
In retrospect it's pretty obvious this was central to his character: he couldn't accept he got hella lucky with dilbert happening to hit pop culture square in the zeitgeist, so he had to adjust his worldview into him being a master wizard that can bend reality to his will, and also everyone else is really stupid for not doing so too, except, it turned out, Trump.
From what I gather there's also a lot of the rationalist high intelligence is being able to manipulate others bordering on mind control ethos in his fiction writing,
Supposedly it reduces std transmissions and thus HPV caused cancers, which of course is neither here nor there with respect to doing it to people who can't meaningfully consent, and also there are non surgical alternatives like condoms and hpv vaccines.
Female circumcision is usually butchery and completely indefensible.
Turns out even the paywalling is fake since you can still do the edits by accessing grok from other parts of the interface like context menus, you just can't outright ask it in a tweet.
heads up I heavily edited the post while you were responding though I don't think the essence changed. I added their reasoning for only allowing image generation for paying subs and expanded the table a bit.
In response to the outcry image generation is now turned off for non-paying users. Ostensibly it's so people who are not using Grok as god intended can be identified via their subscription information, but I can't help but think it's just Elon explicitly monetizing grok's CSAM capabilities.
It's not just the anglo governments dropping the ball, there's also the various app stores who simply won't enforce their own explicit rules and ban the Grok app, I guess because the notion of accountability for the effects of AI slop must remain unthinkable for as long as possible.
If the great AI swindle has taught us anything, is that what's good for normal people isn't really important when all the macro-economic incentives point the other way and towards the pockets of the ultra rich.
As of April 2025, only 17% of Americans thought AI would have a positive effect on the US over the next 20 years. Only 23% thought AI would be positive for how people do their jobs.
An important takeaway I think is that instead of Actually Indian it's more like Actually a series rushed scriptjobs - they seem to be trying hard to not let the llm do technical work itself.
Also, it seems their sandboxing amounts to filtering paths that star with /.
The not-everyone-agrees link is to acx and siskind's take on the matter, who unfortunately seems to continue to fly under the radar as a disingenuous eugenicist shitweasel with a long-term project of using his platform to sane-wash gutter racists who pretend at doing science.
The common clay of the new west:
Twitter post from @BenjaminDEKR
“OpenClaw is interesting, but will also drain your wallet if you aren't careful. Last night around midnight I loaded my Anthropic API account with $20, then went to bed. When I woke up, my Anthropic balance was $O. Opus was checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night." Doing literally nothing, OpenClaw spent the entire balance. How? The "Heartbeat" cron job, even though literally the only thing I had going was one silly reminder, ("remind me tomorrow to get milk")”
Continuation of twitter post
“1. Sent ~120,000 tokens of context to Opus 4.5 2. Opus read HEARTBEAT md, thought about reminders 3. Replied "HEARTBEAT_OK" 4. Cost: ~$0.75 per heartbeat (cache writes) The damage: