开启时会模糊预览图，关闭后正常显示
I hear the concerns, so I wanted to be 100% and share what I do with Stripe API keys shared on TrustMRR: 1. When you upload the key, it hits my API endpoint (NextJS /api on Vercel) and gets a few information from Stripe 2. The key needs to read Charges and Subscriptions data to reconstruct your revenue and MRR. It also needs read-only access to Connect and File so I can use your Stripe account name, icon, and description to build a nice page. 3. Once done, I compute this data and store it in my database. I only store what TrustMRR needs to display a page: revenue per day, current MRR, startup name etc. If you don't see it on the page, it's not stored in my database. 4. Your API is then encrypted in my database in MongoDB. That way, every 12 hours, my cron job queries Stripe again, fetches new payments, and updates your revenue data. I just hired a security expert to audit the codebase because I know how important security is. That's it!
For instance, this is everything I store for this startup of mine: https://t.co/wix2POHCuF
RT @KnightNemo_: Introducing🌍 Awesome-World-Models, a one-stop github repo of everything there is to know about world models! Here is a ne…
open-source AI · protocols · programming · paragliding · building https://t.co/Cdhh8ECiyd · previously https://t.co/2fw43SSiR5 · 🇺🇸🏳️🌈
Here is a better walkthrough of the 6million splat scene using sparkjs LoD on a Quest3. Gives a better idea of the size and detail of the scene. Controls are super broken which is why I run into walls :) roughly 30 fps. 500k visible splat budget.
GP @ a16z ... questionable heuristics in a grossly underdetermined world
Taylor Sheridan's New $1B Deal, Explained https://t.co/wQXkylFwr8 // The numbers on this work are incredible, just like the resulting output.
Subscribe https://t.co/Xm1OaUU8jk • seed investing • writing • ॐ •🙏• I use '—' • tweets saved 90 days • 📷
app will be 90% free coming SOOOON, whole new revamp to maximize fun of love and love of fun
curious guy creating things @ https://t.co/HXWladhJaA - future wife guy
docs as skills + “Memory Banks” for portable memory: been thinking about this pattern for a bit. In Agent Eng, we can do the pre-work of bringing docs/config into a centralized place (as Skills). Locally loaded with the agent means less errors from finding the right thing via web search. Also Skills have Progressive Disclosure so we get way less context bloat (may still want to specialize agents with subsets of skills) “The Memory Bank”: treat part of your agent memory as a portable store for Skills. Have a centralized Memory Bank and attach selected subsets to agents as needed (fresh docs, scripts, etc). The Filesystem architecture is so powerful with agents because we dynamically add “functionality as files” and decide how/when it gets exposed to agents really powerful + simple agent pattern, lookin to see/hear how we can help others with docs/capabilities management
building agents and harnesses, prev @awscloud, phd cs @ temple