π“Ÿ host1.bitchan.one

𓀐 @Cee6M7nM

⏱ 1w seenβ›“ 0πŸ‘ 0β§– 1

filter [-]

languages:onother

is there anything opus 5.5 can't do when it comes to programming? i mean most companies are already using it for 90% of tasks with humans only being there for basic preference guiding and extreme edge cases. I feel like we'll start seeing a regression soon. companies like anthropic will put so many safety filters in place that it'll effectively neuter the model because they cannot "waterdown" such a powerful model for the average user. in fact, they're already trying to put those systems in place https://safeguards.anthropic.com/2026/self-assessment/ >Today, we’re introducing self-assessment, a technique that lets Claude provide a safety assessment of its own conversations.⏱ 1w↩ 1
I'd a model that can handle large codebases well, forgets things less, and is generally just smarter and "gets you" more.⏱ 1w↩