AS

@kapilansh_twt You can use Mozilla sops and a key, i.e. gcp Kms or age for offline use

how do teams actually share .env variables securely because the options I see are - Slack DM (terrible) - email (worse) - shared Notion doc (somehow even worse) - 1Password or similar - something I'm missing

I hate the 'You've opened 100 container tabs' firefox bullshit.

@PandelisZ MCP server for billing dashboard when

Navigating to your AI provider's billing dashboard is the hardest task of my day. Im either thrown into a claude window, a wild vertex page, the wrong chatgpt account or stuck in an auth0 flow

@simonw @hxiao @dottxtai I was also under the impression that different providers support different JSON sub schemas. I.e. openai requires that all fields are required, but each field can be nullable, whereas Gemini allows for optional fields but no field can be nullable.

@hxiao @dottxtai Currently the LLM schema mini-language doesn't support optional fields, maybe it should? The problem with required fields is they they can force hallucinations

Highly recommend this post by @simonw about extracting structured data from text with LLMs. Simon puts it well: "the single most commercially valuable application of LLMs is turning unstructured content into structured data." Link below 👇

@nsphin @Davidjpark96 @idanmasas Why did you have to pivot if you don't mind me asking? I've been interested in joining this space for the last year or so

@Davidjpark96 @idanmasas We started an edtech AI startup selling to schools.. and basically had to pivot to not die before the sales cycles end If ever you share any playbook, I am all ears!

Our 2nd product just hit $100k MRR 🥳 It's been super fun re-living these early growth milestones again via acquisitions Will likely be finishing up 2-3 more acquisitions this month! Current ARR of our tools combined: $8.96M

@nisten Have you used SNOMED CT (en.m.wikipedia.org/wiki/SNOMED_CT) or ICD (en.m.wikipedia.org/wiki/ICD-10)? They are medical ontologies that should contain most diagnoses/diseases. Each code can be nested, with each nesting allowing for more specificity on the type of disease (e.g. Diabetes, I Vs II)

There was no easy list of all human diseases & conditions so I just made one. Manually cleaned 2000+ entries. If I missed anything leave a comment or pull request. Will be MIT license after doctors review it. If you want to contribute just contribute here: huggingface.co/datasets/niste…