Such a complicated issue - but here's my response. I will gladly pay you today for access to Fable 5. I will go through any kind of background check and log into a secure system with a fricking security dongle encrypted blood sample give you my firstborn son (sorry, Hayden, it has to be done) to get Fable 5 access back.
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: https://t.co/bwn0sximKZ
@Samaytwt I've been using https://t.co/u3IcytAqJR Max for months now... I work 10 to 12 hour days and have multiple sessions running in PowerShell... have never hit a limit.
@ravikiran_dev7 The ‘coding’ has its own version of tech debt. Especially in enterprise platforms. Even more so are the apps that have been around for years… I’ve seen behind that curtain and oh my gosh!
@jeremyphoward@AnthropicAI Can they make their AI more affordable? I guess. But, when I see posts like this I ask myself, "Self, are they using AI for agentic AI in production or are they just arbitrarily using the API for development?" If it's the latter, maybe that's part of the problem?
@alex_prompter They should use Claude dot AI for R&D and use the API only in production environments. Nobody needs to experiment and iterate on the API.
Any new change happens in a sandbox before being migrated to ykur production is how I’m handling it in my platform. The problem is what the nature of the infection is… if it’s a prompt injection attack it’s handled and reported with prejudice. All work stops. If it’s a code level injection my hosts system should catch it. That’s the theory right now. We’ll see when it goes live.
@bcherny Claude is consistently under-delivering. Stubbing, mocking up, quitting early, deferring... no matter what, I can't make it extend/over-deliver. It just contracts/under-delivers. I'm gonna go watch TV and enjoy some whiskey. Hope you guys are working on this.
I've spent thousands of dollars on Claude - i'm invested. But the rate at which, lately, it stops working, defers on plans I've asked it to build, stubs and mocks things up instead of wiring - I used to call it lazy. Now I'm calling it derelict.
I don't know why I keep paying for Claude Max, it lets me down far too often to justify the loyalty.
Ugh - the absolute worst move. HTML is for humans. LLMs do not need the bloat of HTML. Markdown had its place but it was bad enough. HTML for LLM consumption is worse.
Claude Code users are quietly switching from Markdown to HTML.
Specs, PR reviews, reports, prototypes, editors.
HTML is becoming the new agent artifact.
Bookmark this.
@srishticodes Hey - nice pattern/concept... but this will not work. While it's a reasonable set of preferences and instructions... but that's all they are. There's no enforcement layer, no state machine, no actual gages. It's aspirational at best.