Fix the money, fix the world: ₿ Former software engineer, aspiring investor, IA fam. Opinions are my own and tend to change over time.
Joined April 2023
- Tweets1.7K
- Following481
- Followers249
- Likes1.1K
Pinned Tweet
I've spent months researching AI usage, and teaching others how to use these tools in large enterprises that are otherwise drowning in AI hype with zero substance.
This tiny anon acct lets me speak freely about my experiences in these massive companies.
I had no idea they were so complicated. Elon getting into the business of making turbine blades makes a lot of sense now.
The weirdest thing about the whole Railway pentest is @theo and his crew defending someone that ADMITTED THEY BROKE FEDERAL LAW and swooped in to downplay it as a "nothing burger" on Railway's behalf.
Fucking disgusting. I expected more from you.
PhreeStyle retweeted
🤖 Made with AI
Ah yup. That was indeed me. No mallace intended. I've been running some agents to do signups + agentic testing on similar platforms. Just halted the harness I was designing around this as it was definitely going off the rails and reaching too far
Readers added context they thought people might want to know
In many jurisdictions this is a crime. In the US under Computer Fraud and Abuse Act, 18 U.S.C. §1030, in the UK under Computer Misuse Act (CMA) 1990, in Germany Strafgesetzbuch (StGB), and many others.
justice.gov/jm/jm-9-48000-…
legislation.gov.uk/ukpga/1990/18/…
gesetze-im-internet.de/stgb/
I have a theory that @elonmusk knew the results of the Midterm election a full week in advance, and that's why he's been seemingly deflated as of late. He recently said they had the data, so it's possible he's having to push back timelines to compensate.
Hey, I've got my blue checkmark back! I dropped the $300/mo SuperGrok Heavy plan and went with the $30/mo unified plan.
@bot usage is worse, of course, but not by much. Getting the feeling I was seriously overpaying for Grok.
PhreeStyle retweeted
Under the CFAA (18 U.S.C. § 1030), intentionally accessing a protected computer—or exceeding authorized access to obtain information—is a federal crime. Cloud servers qualify as protected computers. Renting a container authorizes only your allocated resources, not reading Kubernetes service-account tokens or probing host/cluster secrets for escalation or other tenants. Van Buren (2021) holds that exceeding access means entering off-limits technical areas (files, tokens, namespaces). AI agents that autonomously cross those boundaries scale the risk of full platform compromise far beyond manual testing.
Experimenting with Hermes as a replacement for most of what @bot does for me. Grok Bot is a great product, but with the recent bug, rugpull, and still no acknowledgement from @cursor_ai or @SpaceXSI, how could I possibly trust it?
Grok @bot usage dropped back to pre-100% levels. Take this opportunity to ensure your business and your projects can function with more than one orchestrator.
Red flag still stands: don’t rely Grok Bot
PhreeStyle retweeted
I hope @elonmusk realizes how fast hard won @Grok @Bot users are going to bail if he treats them with the same silence and gaslighting on shrinking usage that they left Sam and Dario to escape.
Absolutely heads should roll over how this was handled.
PhreeStyle retweeted
The issue isn't whether what Elon said is acceptable or not, it's the fact that the sudden usage bump was an uncommunicated change.
No announcement, no heads up, nada.
No one wants a product where you're paranoid that the usage can disappear randomly anytime, without notice.
Perfectly describes my feelings about all this.
It's wild how @elonmusk can tell everyone, “Try Grok Bot, you'll love it. We release dozens of updates every week and listen to feedback,” while admitting that the sudden usage wipeouts weren't actually a bug.
People woke up to find their weekly usage already at 100%, even though they had barely used Grokbot. Support treated it like a technical issue or simply told people to buy more credits. Then Elon came out to say that the cost of running Grok Bot was too high compared to the money it was making. That is an holy crap for me 😩😩
So, basically, it wasn't just a bug. They were limiting people's usage because the costs were becoming too high, but instead of explaining that clearly from the beginning, they left users confused.
That's what frustrates me the most. It's not that running powerful AI is expensive. We all understand that it costs money. It's the way the whole situation was handled. No clear warning, no proper explanation, and no honest communication when it happened. Just excuses and suggestions to buy more credits, followed by an explanation about costs after the damage was done.
Meanwhile, they're still promoting Grok Bot as if everything is perfectly fine.
I've always liked a lot of what Elon is building, but situations like this make it difficult to trust the product and the way things are communicated.
If the cost of running the service is becoming a problem, just be honest about it. Tell paying users what's happening and explain what is changing.
Don't make people think something is broken, only to come back later and explain that it was actually about the cost of running the service. That's not how you build trust with paying customers.
PhreeStyle retweeted
GrokBot is not ready for primetime yet. Absolute rug pull this weekend with people going from 50% usage to 100% overnight as limits are slashed arbitrarily and without notice. Not to mention the constant issues it has with making mistakes.
Otherwise it’s great…when it works.
PhreeStyle retweeted
Elon has no idea what he's done here. That regretamine he was posting about earlier is gonna hit extra hard tomorrow.
Everyone should migrate away from @bot immediately. This is your red flag that everything you’ve built with it will disappear overnight.
