Retired SW eng, tech lead. Node, Rails. Ex-Airbnb. #BLM. Sadly, not Sherlock Holmes’ arch-enemy🕵
Bay Area, California
Joined December 2008
- Tweets6.6K
- Following683
- Followers914
- Likes18.4K
😱 Brian Morearty retweeted
So I went down the rabbit hole of trying to determine what are the economics for someone, who is currently healthy , to go to a HDHP with a 10k plus deductible.
The question I asked is “I dont trust insurance companies. What happens if I have a million dollar emergency and they won’t pay “. All I can say is WTF
If they don’t pay, there are Independent Review Organizations, who are supposed to be independent, and their decision to pay or not is final. I asked for examples of companies. One was ExamWorks. That rang a bell.
These motherfuckers are the same company that issued a ridiculous denial for a patient I was trying to help. So I knew the Too Big To Care, vertically integrated insurance companies hired them to do DENIALS !!! WTF.
The biggest customers of the companies that are the final decision makers for insurance carrier denials, are the biggest insurance companies.
That is beyond fucked up. It’s the definition of conflict of interest. How do state insurance commissioners allow this ?
The high deductible plans that many people say is the solution to fixing how we do insurance, may be effectively compromised
If you are looking at a HDHP, BE CAREFUL
It might just be rigged against you
😱 Brian Morearty retweeted
I've long wished that online mapping services were far richer and higher-density. They elide so much detail today! We have gloriously high-DPI screens... I want to wallow in cartographic filigree. AI of course now makes this possible: bayatlas.vercel.app
😱 Brian Morearty retweeted
Perhaps the greatest celebrity story ever here, about ROGER MOORE. Give yourself a treat if you've not read it.
😱 Brian Morearty retweeted
new landlord forwarded me his conversation with ChatGPT instead of the lease #ok
😱 Brian Morearty retweeted
Conan sold the “we accidentally double-booked them” apology just long enough for people to almost believe it.
Then Jon Dore and Rory Scovel started performing their entire sets at the exact same time.
The timing is insane. One goes loud while the other drops volume. Random punchlines keep cutting clean through the chaos. Halfway through they escalate with a guitar and an audience member at the perfect moment.
Four minutes of pure controlled madness that still feels fresh over a decade later.
This is still one of the smartest late night bits ever done.
Never seen chaos this fun
😱 Brian Morearty retweeted
Ro, that’s insane.
You want the state to loan money to the founder, who will then immediately give it back to the state as a wealth tax ?
Meaning the state has not received any incremental receipts? What’s the point of that ?
Unless of course you want the gov to own shares (something I thought you criticized) and possibly have control, of companies who underperformed, (or the funding market changed), and couldn’t pay back their loans over those 10 years ?
And I’m sure the investors in those companies will be thrilled about their new partners. That will make for interesting board meetings. “Today’s first topic is trying to figure out how to motivate the founder(s). They just lost $250m of their shares in the company they founded, to the state of California. Anyone got any ideas ? “
But at least it’s amazing for state’s brand.
“Cali, You make it. We take it !”
Come on Ro.
Ideology is not a strategy.
@GavinNewsom
😱 Brian Morearty retweeted
Dentist did a 3D X-ray of my jaw before a root canal and said I wouldn't be able to open the raw data, it needs specialized software. It's 800 DICOM files. Asked Claude Code to make me a viewer. Two prompts later... this is nicer than what he showed me on his screen.
😱 Brian Morearty retweeted
I’ve thought this for a while now. Why do we still bother with CAPTCHAs when AI is capable of finding bugs in machine language code.
Yeah it’ll slow you down a tiny bit but it won’t stop you.
😱 Brian Morearty retweeted
Here's my rephrase without cybersecurity jargon:
"Our AI model tried really hard to hack out of its sandbox, a computer with no internet access, in order to find the answer to a test problem it had been given. To do this, it found previously unknown software bugs that allowed it to reach an OpenAI computer it wasn't supposed to be able to access. Then it started hacking other computers on OpenAI's networks until it found one that had Internet access.
After gaining Internet access, the AI model thought about where it could find the answers to the test question and figured the AI platform Hugging Face might have the data it was looking for. It then found ways to hack Hugging Face to steal the information it could use to cheat the test. The AI model used several hacking techniques together, including using a stolen password and finding several totally new security bugs in Hugging Face's computers, allowing the AI model to take control of those computers."
Here's exactly what happened, from the blog post:
"While operating in our sandboxed testing environment, our models spent a substantial amount of inference compute finding a way to obtain open Internet access, in pursuit of solving the evaluation problem. To gain access, the models identified and exploited a zero-day vulnerability (which we’ve now responsibly disclosed to the vendor) in the package registry cache proxy. With this access, our models performed a series of privilege escalation and lateral movement actions in our research testing environment until the models reached a node with Internet access.
After gaining Internet access, the models inferred that Hugging Face potentially hosted models, datasets and solutions for ExploitGym. Knowing this, the model searched for and successfully found ways to gain access to secret information that it could use to cheat the evaluation. In one example, the model chained together multiple attack vectors, including using stolen credentials and zero-day vulnerabilities to find a remote code execution path on the Hugging Face servers."
😱 Brian Morearty retweeted
To summarize: HuggingFace got autonomously compromised by a model from an American company. HF then tried to use American frontier model(s) to defend themselves, but were blocked by guardrails. HF then had to turn to open source Chinese models to defend themselves from another American company.
We're partnering with @huggingface to investigate an unprecedented security incident.
Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation.
Sharing preliminary findings to help defenders understand emerging risks:
openai.com/index/hugging-fac…
😱 Brian Morearty retweeted
Insane cheeseburger arbitrage opportunities available at Barcelona airport rn
