For anyone reading in future, whilst I'm very much anti-unethical AI in terms of training and ripping off hard-working folks with gen AI, if an MCP or API integration actually worked, I would be interested in seeing how someone runs a local LLM to organise their tasks and make sense of it all from an existing workspace. For this to not be counterproductive, it would need to be Clippy-like suggestions about moving some tasks to certain projects or making them sequential sub-tasks of another. But this would sincerely need to work well (and I don't think a next-word prediction model is going to cut it) or it would be as annoying and mostly useless as Clippy. Would love to be pleasantly surprised, but I don't think it'll materialize as I personally feel LLMs are a dead-end technology for most of the problems they're being thrown at.
You could probably do it without an LLM just by using RegEx for keywords and looking at dates and times when tasks were created, or looking at the archive of tasks that are already complete, looking for patterns in when they're created, what order they're completed in, context etc. The API excites me quite a bit. On Wednesday, 16 July 2025 at 17:30:45 UTC imajeff wrote: > Now, I'm in no way speaking for the MLO team. I have nothing against > anyone trying to make good use of modern AI, I've been studying AI since > the 80's and it's only that I understand the problems apparently better > than those who said basically *we developed it but we don't know how it > works*. > > MLO team is doing it of course. You should look for the recent > conversation on https://groups.google.com/g/mylifeorganized/ > where they have already responding to the question and said they are > working on the Web version of MyLifeOrganized, it was already announced > that will have a good API, but they said that will be the API to support AI > integration. > Fingers crossed, we'll see how it goes > > On Tuesday, July 8, 2025 at 9:25:08 AM UTC-6 Daniel Clark wrote: > >> All I can say is that so far the AI from my Bee Wearable ( >> http://bee.computer) combined (via MCP) with Shortwave email has been >> amazing. It knows about most things I say and write, plus records family >> history etc. I recently realized how valuable this will be after people I >> love pass away - I'll have access to facts about family history I'd be >> unlikely to write down or remember otherwise. And it is good at >> answering questions based on this knowledge base. >> >> The only thing I find annoying is that I now have multiple places with >> TODOs. Hence the suggestion. >> >> >> On Mon, Jul 7, 2025 at 8:20 PM imajeff <[email protected]> wrote: >> >>> >>> 1. you forgot to write (try 1) on your first post, Lol >>> 2. It's not clear how this AI stuff will work considering the >>> unwanted eternal hype where it has never been nearly as helpful as >>> everyone >>> selling something seems to claim... as well as my experience being that >>> AI >>> only works for typical, but it fails for me because I'm not typical. May >>> I >>> point out that MLO is not typical either. >>> >>> >>> On Saturday, July 5, 2025 at 8:18:14 AM UTC-6 Daniel Clark wrote: >>> >>> I’m a dedicated user of My Life Organized (MLO) and love how it keeps my >>> tasks in sync across Android, iOS, Windows, and soon the web. I also rely >>> on Shortwave Email and the Bee AI Wearable, where I’ve been using an MCP >>> Server (via [beemcp on GitHub](https://github.com/OkGoDoIt/beemcp)) to >>> seamlessly integrate my email and Bee. This setup transforms email into an >>> intelligent extension of my workflow, allowing AI to create todos in Bee >>> from emails, reference past decisions, and understand my priorities—all >>> without switching platforms. I’d love to extend this capability to MLO by >>> adding an AI MCP Server, and I believe it could take MLO to the next level. >>> >>> >>> ### What is an AI MCP Server? >>> The Model Context Protocol (MCP) is an open standard from Anthropic that >>> lets AI models connect to apps and data sources through a universal >>> interface. An MCP Server for MLO would expose its todo lists, priorities, >>> and actions to AI assistants (e.g., in email or wearables), enabling them >>> to read and interact with MLO data securely and efficiently. Learn more at >>> Anthropic’s [MCP introduction]( >>> https://www.anthropic.com/news/introducing-the-model-context-protocol). >>> >>> [...] >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "MyLifeOrganized" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To view this discussion visit >>> https://groups.google.com/d/msgid/mylifeorganized/debb6373-9d2c-4cf7-af19-6641a353c25fn%40googlegroups.com >>> >>> <https://groups.google.com/d/msgid/mylifeorganized/debb6373-9d2c-4cf7-af19-6641a353c25fn%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- You received this message because you are subscribed to the Google Groups "MyLifeOrganized" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/mylifeorganized/66a93570-4b21-4271-9d7a-7a899ef9c67fn%40googlegroups.com.
