"I think it will actually be possible to cure most human disease in ~5-10 years, as crazy as it may sound" is a direct quote from Dario Amodei. In his original essay he makes clear that that doesn't mean all those cures will be distributed to all on that time-scale, but I would still interpret this as a wildly ambitious prediction that AI will cure most diseases. He predicts in the same essay that it will be responsible for a 10X speedup in drug development.
As implemented, it strongly violates the principle of least surprise (assuming this is true).
A more sane implementation of this would involve a prompt the first time an agent attempts to connect to each of the sites that Anthropic thinks it necessary to send this email header to, asking if this is OK (and bailing on all requests to that domain if the prompt was rejected).
>” A more sane implementation of this would involve a prompt the first time an…”
but users of this type of software generally want an easy experience and don’t want to be bothered with details, only conveniently retrospectively users demand to be involved in the decisions.
it reminds me of a pattern I see in small business owner managers where they want to simultaneously be consulted about everything, but are never around the business. You can’t have it both ways.
I think this part has been prettymuch fixed in recent releases. (I didn't just get used to it, because when I look back at older output, it still drives me crazy with all of these buzzword tics).
Mandatory military service, too. You have to imagine that the US would have a similarly disproportionately-scaled cyber-warfare industry if it also had mandatory military service and good sorting of capable people into those units.
This is the one case I don't mind it. Suggested responses feel like they cheapen human interaction, but here I'm talking to a robot that really does tend to know what I want next, and also is highly unlikely to be offended by a less-than-heartfelt response.
The suggestions in the box where I put my text push me out of flow. Bad enough it says “want to do X?”, but when it’s in my own text box it screws me up. It’s not more efficient. It actively makes things worse for me.
This seems like a really good example of Base CVSS scores not telling us much on their own.
For the Wordpress RCE (nominally CVSS 9.2), it looks like many standard deployments of WordPress would be affected, barring extra mitigations. But in the case of these Hugo ones (9.3), it looks like very specific circumstances (anti-mitigations, if you will) are needed. E.g. running arbitrary builds of untrusted user content without a sandbox; running it in a GitHub workflow against PRs from untrusted contributors, etc.
Parler was "ripped off the Internet" because it was delisted from app stores and dropped by AWS for violating their ToS due to having (If I remember correctly) approximately zero content moderation, which is not really compatible with the mainstream distribution model they wanted. The owners were not subjected to sanctions by a foreign government.
That part seems to qualify as an unsolved problem. But could anyone have taken the scanned data (or the GIF file) and used it to open a bank account in your friend's name? That seems like the main issue that is genuinely solved by correct implementation of this type of system.
Already today nobody can open a bank account in my name with just a picture of my passport, as the original would be required. My passport doesn't have any of the "eID/CAs/ZKP/PKI", so the question of "what exactly the addition of it solves" remains open.
My national ID card supposedly has some of it, the 17-year olds who want to pass as 18-year olds usually show a doctored gif file of their ID card, with a year of birth one or two years before the actual one; this works in ~98% of the cases.
Actually it does. Biometric passports (and IDs) have a chip which is read via NFC and the information the NFC provides is signed by a CA which is the government that issued the passport. ICAO compiles a database of public keys corresponding to each government (plus countries exchange their public keys via bilateral agreements). Unless somebody is doing purely visual inspection, any time a passport is scanned there's PKI involved to validate if the information is genuine.
Private entities get to use whatever name they want. In this case, there is the option to use the longstanding name that matches the one used by the other neighboring territory. Or to have a wait period after the name is changed unilaterally by executive order, because it'd be silly to change it back and forth. Or to put "a.k.a..." after it. Or to wait for the relevant government to update its own map and geographic names database first (https://edits.nationalmap.gov/apps/gaz-domestic/public/searc... is not yet updated as of now). Or to wait until local signage can be updated so it mostly matches the map (though I don't think Rochester will be in a rush to do this).
reply