BBSs were really fast, cause they were text only on dial up. Connectors or MCPs, allow Agents to talk by data sets not UI. But also restrict unwanted agents from clogging things down. When you have a flood of zombies bots and human bots, you only want to let the human bots in.
Connectors feel like the obvious way to make the internet agentic.
I think theyāre a dead end.
The internet won because you could go anywhere without asking the destination to integrate with your browser first.
Agents should work the same way.
If every service needs its own connector, you havenāt democratized access, youāve created a toll booth in front of every website.
The internet doesnāt need to become API-native for agents to use it. And merchants shouldnāt have to change anything for software to transact with them.
And there goes another week of productivity, But hey I will have the coolest avatars ever for my esp32-s3.
Asked Muse what Gemini API CLI can do better than it and suggest 5 tools to create with it.
My AI agent just got 2x better.
Gemini API CLI merged with my Muse, so it made a video to celebrate!
(one costly celebration :0) So now we have a budget rule.
š¤ Made with AI
This morning I asked my Muse Assistant to give me the task updates in 3 second overviews so I can make a decision at a glance instead of reading a bunch of text. wow did it ever deliver! Use this code to get a ton of free usage for your own Muse Personal Assistant. EDVDHI
If your Muse or any other Bot agent seems to be making up things, just tell it to QA everything before it responds. It is amazing how much better the actual work gets without even needing to check. I guess it is just the threat of QA. :0)
Honest @Muse feedback:
- 50%+ of the source links are broken and unverified
- Lots of random links to archaic irrelevant websites
- I asked for insta + personal website links
- Cannot effectively search @instagram
Step 1 is links are parsed incorrectly:
- "corduroymedia%7CIG" should be "corduroymedia | IG"
- "episodes/15%7CPodcast" should be "episodes/15 | Podcast"
%7C = | in ascii
cc @alexandr_wang
This is why AI should aways assist in the writing not write it for you. AI is an amazing tool but should not replace brain power or it is not beneficial at all.
Why share information if no one will read it in the first place?
Do you remember what you wrote with ChatGPT?
In an MIT study, minutes after writing an essay, 15 of 18 people who used ChatGPT couldn't quote a single sentence of it correctly. Among those who wrote alone: 2 of 18. š§µ
#AIinEducation #EdTech
This is one of the coolest EdTech tools I have see in a while, can we get more things like this, I love the teacher mode.
I built an interactive 3D Solar System Lab š
Realistic planets, live orbits, teacher mode, student challenges, and 8 science modulesāall in one browser experience.
Try it: orbital-atlas-jc.chen1332642ā¦
#EdTech #Astronomy #ThreeJS #AI #Space
I have always hated holiday shopping, so this year I am going to have Muse find all the deals throughout the month for me! I think AI is really going to revolutionize the shopping experience and Amazon is going to miss out.
Here is the prompt and AI I use
muse.ai/join
Redeem code: EDVDHI in the settings and we will both get a ton of free usage.
Prompt:
Find me the best Black Friday deals and other sales going on this month (November 2026) for the things I'm interested in.
Don't start searching yet. First ask me a few questions one at a time like the following.
1. What am I interested in?
2. What's my budget or target price?
3. Wanting to buy now or jsut watching?
4. Any preferred stores to check, or stores to skip?
5. Do I need shipping, local pickup, or in-store only? What's my ZIP/town if local deals matter?
6. How often do I want sales alerts? For example: instantly when something drops below my target, a daily digest, a weekly roundup, or only during Black Friday / Cyber Monday week
7. How long should you keep watching? Until I buy it, until the end of November, or until Cyber Monday?
Etc...
Once I answer, set up that sales-watching task and start checking.
When you report deals, only report prices you've verified on the live retailer page, not just search snippets. For each deal include: the exact product/variant, the sale price and regular price, total with shipping/tax/fees if you can confirm it, the store, when the sale ends, any coupon/code needed, and a link to the product page. Rank the best deals first, flag anything that's out of stock, a different variant, or a resale listing, and don't buy anything without my approval ā just alert me.
Having an AI Agent run my Machine learning training and evaluations has been wonderful. Embedding Gemma and Hunch has helped improve, and now after getting to 87% recognition it made me a portal to manually approve ASL signs it cannot id.
Looking forward to integrate Gemini 4 Carbon's 3d ability to the development of 3d pattern recognition.
It has been best for reviewing and comparing results for multiple models.
Been working on a Machine learning too that takes ASL video creates 3d wireframes of the sign movements so we can get better representation of words at different angles. This afternoon gave Codex a goal and stepped away, came back and it said it is almost at 100%, dug deeper, it took away the 3d vector movement and went back to video...
Using Luna to automated multiple ASL Machine Learning training, just realized it was not using my GPU. Got it fixed and wow much faster. :0)
Luna is great at some things but miss over bigger pictures like larger models.
I haven't bought a new video game in years, the new Mass Effect was what I would have spent money on. Can't wait to see more of these.
If they take Mass Effect from us fans, weāll find a way forward.
Hereās what I can make in a few hours. Two missions and multiplayer coop already working. It wonāt be Mass Effect but what we love doesnāt need the brand to keep going.
You canāt kill an idea.
#SaveMassEffect
youtu.be/W-BtNiUigtc?is=Zwcbā¦
Don't be like a chat bot and think theoretically, be like an agent and just do it.
She inspired me to just do it earlier this week. Yesterday had my first call with a potential client on AI automation of their marketing.
44% of cold WhatsApp messages got a reply. In 5 hours. On a Saturday.
I messaged 18 Croatian webshops about a Christmas advent calendar campaign:
š¬ 8 replied
š 3 interested
š 1 asked for an offer
The secret: no pitch. Each first message had a personal compliment about their shop and one question: āAre you planning any Christmas discounts?ā
I'll send another 20 on Monday!
P.S. Message says: Sounds interesting. Send us an offer.





















