DavidPP

Quick correction, from what I understand Dax and Adam started from scratch for the new version.

I'm currently building an internal tool using SurrealDB directly, but I'm curious to use Morphik since it implement features I hadn't the time to figure out yet. (For example, I started with hardcoded schemas and I like how you support both).

Minor nitpick, but the README for your ui-component project under ee says:

"License This project is part of Morphik and is licensed under the MIT License."

However, your ee folder has an "enterprise" license, not the MIT license.

In which client and with which LLM are you using it?

I use it in Claude Desktop for the right use case, it's much better than thinking mode.

But, I admit, I haven't tried it in Cursor or with other LLMs yet.

Are you using Typesense for the search server? If so, are you using hybrid search with embeddings, or just regular search?

Goose from Block is also a client that supports MCP. It can be used both with UI or as a CLI: https://block.github.io/goose/

On my end, I use https://github.com/skydeckai/mcp-server-aidd, which has

- File manipulation - Directory manipulation - tree-sitter integration

and more.

I also installed Tavily Search, sequential thinking, and Playwright.

I still use Cursor for development, and I use Claude Desktop for higher-level documentation, testing, etc.

For example, I'll check out a new repo that is lacking in documentation. I'll get the app running, then explain to Claude where the code lives, how to access the real app, and how I want the features documented.

Then Claude will happily scan the codebase, take screenshots of the running app, etc., all by himself, and then create a report (through the artifact system) with visualizations, graphs, etc.

I'm new to the world of graph, and I just started building with SurrealDB in embedded mode.

If you don't mind taking a few minutes, what are the main reasons to use Kuzu instead?

Haven't had time to try it out, but I've built myself a tool to tag my bookmarks and it uses 3.5 Haiku. Here is what it said about the official article content:

I apologize, but the URL and page description you provided appear to be fictional. There is no current announcement of a Claude 3.7 Sonnet model on Anthropic's website. The most recent Claude 3 models are Claude 3 Haiku, Sonnet, and Opus, released in March 2024. I cannot generate a description for a non-existent product announcement.

I appreciate their stance on safety, but that still made me laugh.

Thanks for taking the time to explain your process. I'm just starting out, and so far I was planning to:

- Radar: just cool categorized articles, libraries, apps, etc. that I found during the week and organized in weekly drops.

- Guides: In my mind, those were closer to TIL than real guides, though I couldn't figure out how to name them. I love TIL for that.

- Playground: Forcing myself to share prototypes, snippets, etc.

I also have tons of LLM chats that could become content. Instead of trying to rewrite them, I will just share them as they are (while being transparent).

The main idea, as you said, is to find a modality that works well for me and force myself to write more.

1. My personal website. -> I spent so much time ingesting content, I finally realized I need to produce some too.

2. A AI assisted brief generator -> clients often have a hard time articulating their requirements for new projects.

3. Prototyping the UX of "my" version of the perfect "process aware" editor. More organized than a Wiki, more flexible than tools like Jira, Aha and all. Not ready to share a public link yet. My goal is share my mockup in a week or two.

We use https://typesense.org/ for regular search, but it now has support for doing hybrid search, curious if anyone has tried it yet?

Thanks for the link, wasn't aware of this solution.

We currently using https://refine.dev/ with NextJS and Material UI for a CRUD type project. How would you compare yourself to Refine?

Same here situation here (went from paid to free for a similar reason). We are a small team, so we don't represent a lot of $, but it's changed how I perceive Gitlab as a company, and it also means I won't recommend them anymore to clients.

I've been playing around with https://directus.io/ the last few weeks.

I'm really impressed so far, it can either act as the only backend, or as an admin tool over an existing DB. It also has built-in support for "flows" to automate support tasks.

Funny, I was asking myself the same question this morning, I was wondering if there was the equivalent of SETI@Home but to pool resources for model training.

Node-RED is also fairly popular in home automation circles, but it's a generic tool that can be used for different purposes.

For example, it can also be used to build chatbots.

https://nodered.org/

Subscriptions are why I'm torn on the whole Apple Payment system.

On one hand, I disagree with the lost of liberty, the inability to install what I want without Apple approving everything (and taking a huge cut).

On the other hand, I've had so many subscriptions that I forgot, or that make it "hard"to unsubscribe.

That's why I now try to subscribe through Apple whenever I can so that I can have a single view of my subscriptions, when there trial end, etc.

From a UX, perspective, Nielsen Group is a good reference: https://www.nngroup.com/articles/infinite-scrolling/

For your question specifically, the following explain it well I think:

..locating a previously found item on an extremely long page is inefficient, especially if that item is placed many scrolling segments down. It’s much easier for people to remember that the item is on page 3 than it is to gauge where the item is positioned on an extremely long page.

You made me check only me end, and indeed Control Center was sitting at 5.19gb.

I saw this post on Reddit 2 days ago : https://www.reddit.com/r/TeslaLounge/comments/j59ovh/so_tesl...

He was told to Tweet about it and that media would pick it up, sounds like it worked.

I find that Integromat strike a good balance between ease of use (vs pure code) and "power" (compared to Zapier et and IFTTT)

Been using https://www.dendron.so/ for a total of ... 12h.

It's been mentionned here a couple of days ago actually. It's built on top of Vscode and use pure md files, so can be synced with Git or be built as github page.

I think OP meant Anonymous as it was used when FTP was still a thing where it simply meant "without being authenticated as a named" user.

Do you have somewhere we can register/watch to be notified ? Also interested here.

I personally use Integromat which I find is a nice "in between" something simpler like Zapier and using code.

The ability to run code in my own infra is definitely a plus though, will give it a try.

I interviews interns from a school where they all ask their students to rates themselves like that.

At first I was put off, but now I interpret it as their own personal scale, with 5 being what they the best at, and not as "I know everything about this".

But even then, the best candidates are usually the ones who rated themselves the lowest.

Haven't really used it yet, but https://youneedawiki.com/ seem like a good hybrid approach, using Gdoc for editing, but having a browsable read version of your wiki.

He since said sorry on Twitter: https://twitter.com/rus/status/933144474263339009

I meant specifically when building a CMS. Being able to build a CMS that is multilingual, that can adapt to any form of content and workflow + be easy to use for editors is much harder than implementing the underlying technology.