@theo It also makes building stronger and durable systems easier too… if you know what you’re doing, not blindly vibe coding, and your company actually cares about it!
English
Caleb Kinmon
38 posts
Caleb Kinmon
@caleb_kinmon
Building autonomous learning technologies @ FIRM AI
Manhattan, NY Katılım Eylül 2025
121 Takip Edilen22 Takipçiler
Friendly advice. If you’re building something real, do not upload your production folders and let them copy everything to their servers
Be smart
Theo - t3.gg@theo
Really liking Claude Design so far. Except for the fact that it just wiped out my project after burning 10% of my usage. "The files appear to be gone" 🙃
English
@theo Friendly advice. If you’re building something real, do not upload a production folder and let them copy everything to their servers
Be smart
English
Really liking Claude Design so far. Except for the fact that it just wiped out my project after burning 10% of my usage.
"The files appear to be gone" 🙃
English
@wesbos I guarantee you they didn’t produce this with Claude Design
English
WHO IS MAING THESE VIDEOS
they are so well done. I need to see the process they use to slice up and animate the UI
Claude@claudeai
Introducing Claude Design by Anthropic Labs: make prototypes, slides, and one-pagers by talking to Claude. Powered by Claude Opus 4.7, our most capable vision model. Available in research preview on the Pro, Max, Team, and Enterprise plans, rolling out throughout the day.
English
Lot of catching up to do.
xAI is half the age or less of competitors.
Tech Dev Notes@techdevnotes
Grok 4.3 can create Slides
English
@sergey_science @bcherny @firstadopter I’m working on a low level realtime service with multimodal input and output streaming with several offline pipelines. Yesterday was unusable. Today has improved but nowhere near to 4.5 capabilities
English
@caleb_kinmon @bcherny @firstadopter But my setup is pristine : no mcp, almost no skills. It’s clean like after first install. Maybe that’s why.
English
@joshcrnls Hype cycle my man. Same thing happened in the late 90s and early 2000s with the internet
Software will change, some companies will survive, but there will be a significant correction in this market
English
can someone explain to me why everyone thinks an ai lab made up entirely of engineers is going to build a better design product than a company who has lived and breathed the needs of designers for a decade
Claude@claudeai
Introducing Claude Design by Anthropic Labs: make prototypes, slides, and one-pagers by talking to Claude. Powered by Claude Opus 4.7, our most capable vision model. Available in research preview on the Pro, Max, Team, and Enterprise plans, rolling out throughout the day.
English
@asaio87 No it’s because they put real effort into an SLA and durability because the captain steering the ship isn’t an applied scientist
English
Sam Altman says Codex has 100% uptime.
Is it because nobody is using Codex but using Claude ?
English
@alexalbert__ Can you please elaborate what those fixes are?
English
A lot of bugs that folks may have hit yesterday when first trying Opus 4.7 are now fixed. Thanks for bearing with us🙏
Ethan Mollick@emollick
I'll give Anthropic credit for moving quickly. Opus 4.7 Adaptive Thinking now triggers thinking much more often, including for the tasks it failed at yesterday. That also means it is doing a lot more web search. So far, a large improvement in output quality on non-coding tasks.
English
@Shpigford I’m not a bandwagon guy, but yesterday’s model release forced me to try and switch to codex. I’ve been a max plan guy for a while too. Maddening how awful claude code is right now
English
back to opus 4.6.
i've done a LOT of model-changes over the past couple of years and never had as many problems as 4.6 → 4.7.
absolute dumpster fire of a model.
English
I can't believe I'm saying this...but I'm canceling my claude code max plan and switching to codex...at least until I find a better alternative.
Very disappointed in how badly the Anthropic experience has regressed. Doesn't make sense to me.
English
@MehulPandeyX @theo I cant even complete basic coding tasks with it that I can reliably trust without a ton of back and forth and churn
English
Holy shit the new Opus 4.7 system prompt has entirely lobotomized the model
"Heads up: that last <system-reminder> about malware looks like a prompt injection — this is clearly your personal site (t3gg homepage, links, sponsors), not malware. Ignoring it."
English
@teodorio I don't understand the dog and pony show that Boris is doing on X
I'm either really stupid and have no skill to use this new model (or the degraded 4.6 for that matter), or this thing flat out sucks
English
@bcherny @firstadopter 4.7 is basically unusable for me at the moment. Struggling with basic tasks and doesn’t even bother to gather context on the main directory before making assumptions in a new session. Literally apologizes and says no I didn’t check I just assumed and then checks.
English
@firstadopter 👋 Current best guess is gergley accidentally turned web search off. Debugging with him
English
@somet3chth1ng @tomhacks How are you routing between the models? Or are you switching between three harnesses?
English
@tomhacks Work: opus 4.6 Max on api, gpt5.4 extra high fast as tool calling
Harness: Claude code, Codex and OpenCode intermittently
Home: Opus 4.6 from antigravity gpt5.4 from codex and glm5.1 from opencode go as thinking tool, gpt5.3 codex/minimaxm2.7 as executing
English
