This is completely wrong. T3 code is making the Claude Code experience better and enabling multi tasking which increases usage. This should be a net positive for Anthropic to move people up from $20 or $100 plans to the $200 Max plan.
Problem is, retail are a money suck for Anthropic. They use subsidized inference and cost Anthropic a lot of compute.
Pushing retail off of Claude frees up compute for their enterprise customers which is basically the only way Anthropic makes any money right now.
@nate_build@theo He pays $40k a month to Anthropic for his company, T3 Chat. They use the Anthropic API for their service.
He’s not using $40k/month of usage for writing code personally.
It’s to get people to use their inference, not their apps. The inference is where they make (or lose with subsidies) money.
I’d be willing to bet Anthropic lose money on retail subs. All of their profit is in enterprise.
If retail is taking compute away from enterprise, they have a problem. Easily solved by alienating the power users that cost them the most and freeing up compute for enterprise.
@theo wow, so now I have to either give up using t3 code (not gonna happen), or switch to GPT and deal with the sub par UI? Why are Anthropic just shooting themselves in the foot on a daily basis?
@theo On a side note @theo it would be nice if t3code could auto downscale images to fit within Opus 4.7’s new size limit so this doesn’t happen. Seen as we both know Anthropic ain’t fixing this anytime soon…
Ran into this interesting bug. Upload one image that’s over their dimension limit and it nukes the entire conversation and you lose all context rather than just rejecting the image and not adding it to history.
I literally just screenshotted the hero section of my site and it broke. Good old Anthropic, missing edge cases just for fun. Where’s the QA? And is this even an edge case?
@theo any way to use Opus 4.7 on t3code right now? Having to go back to the extension, CLI or, god forbid, the Claude desktop app is too painful after using t3code for the last week or so.
The world’s top trading firms, banks, and HFTs rely on @BSONetwork's ultra-low-latency global network. Very few crypto protocols have access to this infrastructure.
@tier1_x is built directly on BSO's network.
This marks a DeFi industry first: purpose-built, ultra-low-latency connectivity deployed at the foundation of a blockchain.
Long trusted by traditional markets, this infrastructure now powers our chain, the execution layer behind Tier 1.
This is how Tier 1 brings CEX-grade, deterministic execution standards to DeFi.
Fair, predictable & deterministic execution on every block.
No delays. No reorders. No MEV.
Tier 1 enforces arrival-based sequencing for all participants.
Ordering cannot be manipulated or reprioritized.
Execution reflects true arrival time.