@martechismktg had a piece last week about an insurance content team that bought into AI, volume jumped, and delivery did not get faster.
The approver in the story is named Claudette (Real person, ironically). She had five times as much to sign off.
She became a worse bottleneck than the old writing queue.
Faster Gen fills the inbox, but if its not done with proper load mapping accross all affected departments/teams, gains become losses
"Unclear approval chains "
"Disagreements around brand standards"
"Nobody owns the analytics"
If the week got slower after you bought a faster model, open the approver's calendar first, then check on the brand context the model was trained on. And above all, Don't blindly trust AI outputs unless you've proven your specific system
I mean I got that lol especially when its not actually sending any traffic to claude. But my point is that making a game to make fun of something is fine, yet i can't find anything actually giving real examples of Opus degrading that isn't clickbait lol. Seriously I just want examples, I want to be in the know because my experience is the total opposite. every release has been better and better from my perspective
@binghott@digitalnoah So i ask again, for anyone, please give me actual examples of Opus being infuriating and i can almost definitely prove its a process, session, context and skill issue rather than Opus itself
@binghott@digitalnoah See I don't trust this because i dont get to see whats actually happening. If i Download the site as html and then feed it to a bare claude code session it does it perfectly, which further pushes my understanding that this is bias
Amazon blocked Meta's Muse from shopping on Amazon.
Muse launched Sep 8. Personal agent, its own browser, "does the work." Today Amazon said no.
DTC lives on those two platforms. I do not need a future-of-commerce essay.
I need to know if the traffic in the account is a person or a bot with a cart.
No it isn't. Seriously. I keep seeing this sentiment and have been for months, its some weird false bias. I average 20B tokens/month and even when it was 4.5, Opus was used almost exclusively. there is no degradation that I've seen unless you aren't setting up your project properly
@AnthropicAI If I own the repo i'm having an audit done on, with clear chain of ownership, why is there even the option to flag safeguards when doing a security audit. whats the point of even allowing security audits in the first place?
@AnthropicAI Question regarding the Text watermark being rolled out for your models.
In this scenario;
Claude writes a video script, which would have the watermark
If a video is recorded with that exact script ver batim, and then the video itself is pulled as a transcript;
would that transcript also have the watermark?
@zdogmode I'm doing a lot for my current employer but personally, I'm building a DnD Campaign manager because i was tired of every other platform only having a handful of the things I actually wanted, and I'm going to sell it to undercut everyone else since its almost no overhead to me
@zdogmode I'm doing a lot for my current employer but personally, I'm building a DnD Campaign manager because i was tired of every other platform only having a handful of the things I actually wanted, and I'm going to sell it to undercut everyone else since its almost no overhead to me
@zdogmode I'm doing a lot for my current employer but personally, I'm building a DnD Campaign manager because i was tired of every other platform only having a handful of the things I actually wanted, and I'm going to sell it to undercut everyone else since its almost no overhead to me