Leading research at @arcee_ai | Formerly Data Research Lead @DbrxMosaicAI | Visiting Researcher @meta | Ph.D | #TXSTFOOTBALL fan | https://xitter.cf/t.co/4G6Jf3b0V4
Redwood City, CA
Joined August 2011
- Tweets22.6K
- Following2K
- Followers7.9K
- Likes50.7K
No it’s explicit. The way gradschool worked to build a career and a reputation was write papers. Papers are not as important anymore.
Cody Blakeney retweeted
to a rough approximation, 0% of people who complain about ai writing have written anything this insightful or valuable
bj mf it occurred to me upon waking that the west has somehow managed to simultaneously kill the classical and romantic spirits and this explains a lot about what's going on for those of you who are paying attention
hashing this out with gpt and this essay is full of bangers
i would add further that all of this has all happened because men have forgotten god in a solzhenitsynian sense
I’ll just leave this here
Like for fucks sake. The dude has written 10 blogs on his personal website over the past 8 years and as far as I know he didn't tweet about it or hype it or anything. I stumbled upon it and thought the message was cool. Maybe he used an LLM to help him write it, but this isn't AI Slope SEO content farming. Relax who cares.
Why don’t you just @ him @Tim_Dettmers I didn’t write it I just like the sentiment. You are the one complaining on the internet.
I swear to go people saying things are AI slop without reading the substance as if that is a trump card that invalidated any conversation is so fucking annoying. It doesn’t make you sound smart, it actually makes you sound lazy.
Yes there is real slop in the world, yes it’s annoying. Being smart enough to identify that someone has used an LLM to edit writing doesn’t mean you don’t have to read or think critically.
FWIW I’m not talking about my own writing (I make too many spelling mistakes to be accused of slop)
Well you can detect behavior that isn’t human like, and I assume that is happening to some extent. Idk man I’m just responding to the news.
My tinfoil hat conspiracy about this, is it’s not about money or whatever, it’s about preventing agents under the guise of legitimate business to reverse engineer services, products, or probe security.
The yelp one was kind of telling. Agents can do all sorts of stuff if given human level access to a site. Besides potentially allowing customers to make mistakes … I imagine it’s the other things the agents might do that is concerning.
But you can’t just say no to the way people want to be your customer.
Major sites are now blocking AI agents from booking flights and touching accounts. Here's what you need to know.
The story comes from a post by Wise (@trikcode) on Oct 18, which passed 838.8K views. It says Delta is blocking AI agents from booking flights. JPMorgan Chase and Charles Schwab block agents from accessing accounts. Yelp requires paid data licensing for agents. Amazon has blocked Grok bot and other shopping agents. This follows the Amazon block we covered on Oct 15. It is now a wider pattern across travel, banking and reviews.
Banks are the biggest part. JPMorgan CEO Jamie Dimon said outside agents like Muse and Instinct will not be allowed to log in or act on customer accounts. Schwab reportedly blocks tens to twenties of attempts by major AI agents each week. Posts also say Citi and Bank of America now block outside agents from logging in or moving money. Security is the stated reason. Limited read-only access reportedly stays, and moving money is the line.
Travel and reviews are next. Yelp told TechCrunch it only lets in AI agents that pay for its data licensing program. Delta does not allow agents to book tickets. Meta has said turning away a personal agent means turning away the customer behind it, after sites turned away its agent Muse.
Reactions give the reasons. Some say agent traffic is 10x to 50x human traffic, so sites carry the cost. Others point to liability when an agent books by mistake, or to agents holding cards and logging in as humans. Some say banks fear agents moving idle cash into 3%+ accounts. Others expect sites to charge for agent access, or to build their own bots.
Key numbers:
- 838.8K views on the original post
- 4 named blocks in it: Delta, JPMorgan and Schwab, Yelp, Amazon
- 10x to 50x: claimed agent traffic versus human traffic
- Citi and Bank of America reportedly joined JPMorgan
So far the pushback is spreading from retail to airlines and banks.
I can certainly see the need to avoid a constant DDoS from automation, but it seems like the right way to go about this is to make a correct on ramp with like user authentication, so you know who’s agent it is and that you would reasonably expect that person to be trying to conduct legitimate business.
You know like how the whole rest of the site works.
Like for fucks sake. The dude has written 10 blogs on his personal website over the past 8 years and as far as I know he didn't tweet about it or hype it or anything. I stumbled upon it and thought the message was cool. Maybe he used an LLM to help him write it, but this isn't AI Slope SEO content farming. Relax who cares.
Dude who gives a shit. It’s good advice just read the text instead of trying to win internet points.