🇯🇲

177.3K posts

🇯🇲 banner
🇯🇲

🇯🇲

@MidasTouchmkt

I know stuff.. and look at charts 🔊 Two Spirit🪶

Its me vs me 📚4/80 Katılım Temmuz 2009
891 Takip Edilen2.5K Takipçiler
🇯🇲
🇯🇲@MidasTouchmkt·
@zarazhangrui Ehh doing more necessarily doesn't mean actually getting anything done especially if you lack focus or direction
English
0
0
0
3
Zara Zhang
Zara Zhang@zarazhangrui·
My single biggest pain point right now: AI-induced attention deficit I have 5 Claude Code sessions running 10 Terminal tabs open 50 browser tabs open 100 X articles in my bookmarks When there are multiple AIs working for you, you're constantly task switching When you're waiting for AI to output, you're constantly distracted Because what used to be hard is now so easy, you always wanna do more, more, more Because of the speed of the AI news cycle, you feel a constant need to be on this app Add all this up and the result: you live in a constant state of distraction & delirium. Focus & deep work is harder than ever If anyone has figured this out pls teach me how
English
244
33
674
39.6K
🇯🇲
🇯🇲@MidasTouchmkt·
This was my original design idea to visually show what is happening in the memory db for the agent as db grows. As you watch early in the video, the lines connecting the different coordinate spaces have geometry to it as the lines cross and intersect. Now that I have added a 4 dimensional field to this, a thread identifier which is time constraints, those lines connecting become traceable over time no matter how much it grows. It could be 3 years of using the memory db, a singular task can be tracked from the third ever session that happened with the agent. The last image is what happens when data becomes rich in detail; it just looks like noise. I now have a solution for that 😋 the geometrical patterns I mentioned before is the hint lol. Users won't be able to just view the memory but interact with every single thread across the entire spatial domain. I literally wish yall knew how I excited I am about this 😂
🇯🇲 tweet media
🇯🇲@MidasTouchmkt

I can’t wait to create a visually dashboard of my memory db it’s gonna fcking look out of this world

English
0
0
0
56
🇯🇲
🇯🇲@MidasTouchmkt·
@DegenerateNews Anything better than opus 4.6 needs to be Shutdown we can literally stop here and it would be more than enough
English
0
0
0
14
DEGEN NEWS
DEGEN NEWS@DegenerateNews·
NEW: AI COMPANY ANTHROPIC IS TESTING A PREVIOUSLY UNDISCLOSED AI MODEL CALLED ‘MYTHOS’ THAT IS SIGNIFICANTLY MORE CAPABLE THAN ANYTHING IT HAS PREVIOUSLY BUILT - FORTUNE SOURCE: fortune.com/2026/03/26/ant…
DEGEN NEWS tweet media
English
28
9
173
18.8K
🇯🇲
🇯🇲@MidasTouchmkt·
@sabosugi you just made my life 10x easier thank you so much
English
0
0
1
11
🇯🇲
🇯🇲@MidasTouchmkt·
I have a side project where users mainly(woman) who read smut books can have a voice clone of whoever they want reading to them bf, husband ,crush 😂 not only that, they can upload a playlist of songs/ soundtracks from spotify or apple music that will slowly fade in and start to play when the system detects a spicy moment in the book is about to happen 😂 Get a real immersive experience😌 still in development should be ready by summer hopefully
KNOX@knoxtwts

found someone charging $9/month for personalized ai bedtime stories. 900 subscribers. pulling $8k+ monthly. the content is simple. ai generates a 3-5 minute story starring the subscriber's kid. kid's name, kid's favorite animal, kid's favorite color woven into a new story every night delivered via email. this person picked a niche where "made by ai" is a feature. parents don't care if ai wrote the story. they care that their 4 year old heard their name 8 times and is now asking for the doggy astronaut story again tomorrow. most people chase niches where ai tries to replace human quality. beauty content, fitness coaching, educational explainers. niches where the audience actively hunts for signs of fakeness and punishes it. the money moves are in niches where ai capability IS the product. personalization at scale. custom content per user. things a human literally couldn't deliver because writing 900 unique bedtime stories per night is physically impossible. other niches with this same dynamic: custom ai pet portraits from submitted photos, personalized workout plans based on body type questionnaires, ai-narrated family audiobooks from old letters and journals. the product is the personalization. ai is the only way it can exist at that price point.

English
0
0
0
87
🇯🇲
🇯🇲@MidasTouchmkt·
I'm actually more terrified of this actually working and being able to scale, like it shouldn't work lol... but if it does and I show proof this account will probably disappear😅
English
0
0
0
26
🇯🇲 retweetledi
B
B@QuantumTumbler·
You can evolve every possible outcome perfectly. But reality still gives you one. That’s the part physics doesn’t fully explain. Here’s the cleanest breakdown you’ll read ↓
B@QuantumTumbler

x.com/i/article/2037…

English
8
9
51
5.1K
🇯🇲
🇯🇲@MidasTouchmkt·
For example think of a time-lapse of an natural environment changing over time. A pond can expand into a Lake, a hill into a mountain. The Agents memory becomes this, they're are several events and factors happening before the pond becomes a lake. Now the Agent cant navigate each time constraint in this flow outcome in accordance to the coordinates to not only be aware but understand where to go next based on the desired outcome. So as the pond becomes a lake in this time-lapse I can manipulate (desired outcome) the lake to be a certain size, depth etc.
English
0
0
0
23
🇯🇲
🇯🇲@MidasTouchmkt·
I'm basically expanding my 3 dimensional topology memory db into the 4 dimension. For planning the agent will use a numerical number as a time constraint. This provides awareness for the agent while its navigating the coordinates which are events in a multi step process. The constraint also is a connective issue to keep track of what has happened, what is going to happen and what could of been. This happens at every layer for every step in the planning process. Recall happens at correlating constraints and coordinates to provide better optimum routes or a cold start to continue a plan or task that was halted. You don't need to spawn multiple agents for different tasks it just becomes a stream of conscious that just does. Well that is the idea btw lol
English
1
0
0
25
🇯🇲
🇯🇲@MidasTouchmkt·
I can’t wait to create a visually dashboard of my memory db it’s gonna fcking look out of this world
🇯🇲 tweet media
English
0
0
0
77
🇯🇲
🇯🇲@MidasTouchmkt·
What Im bout to create next is gonna make parralel execution by agents look like a joke
English
1
0
0
36
Watcher.Guru
Watcher.Guru@WatcherGuru·
JUST IN: Goldman Sachs says crypto prices may have bottomed.
English
1K
893
9.2K
1.1M
🇯🇲
🇯🇲@MidasTouchmkt·
@ohnahji Eh I don't think you need it tbh, your usage should only scale to the size of what you're building, and sometimes you gotta scale down so things are more efficient if that makes sense.
English
0
1
1
22
Ohnahji 🎓
Ohnahji 🎓@ohnahji·
Let’s goooo!!! Love to hear about multipage long mds and xmls!!! Sheeesh when I get the big bag soon maybe I’ll cop the business plan thing 😂
🇯🇲@MidasTouchmkt

@ohnahji I don't have prompts, I got 3 page Spec MD files 😂 if only you could donate your usage. I hit my limit every Wednesday evening.

English
1
0
1
23
🇯🇲
🇯🇲@MidasTouchmkt·
@_PhallonX0X0 Won' see Elon post about this with "WoW" its really insane how they are treated even admitting guilt to a crime.
English
0
0
1
248
🇯🇲
🇯🇲@MidasTouchmkt·
Look at the progression and listen to me rant about my project 😂you got nothing else better to do. Would love any feedback, also anyone who is currently building stuff feel free to reach out and connect.
🇯🇲@MidasTouchmkt

I wanted a Siri on my desktop to carry out tasks instead of clicking and typing all the time. So I built an interface. It will be able to listen to my commands and carry out these tasks. Lowkey think I 👨‍🍳with the design but now I can sleep😅I had to finish this say hi to IRIS

English
0
0
0
39
🇯🇲
🇯🇲@MidasTouchmkt·
My bootstrap script that takes care of my persistent memory and semantic memory is goated. No matter the model I use, the coding gent can pick up everything from the last session near a 100-token prose without reading any previous files or conversation thread to gather context to begin working.
English
0
0
0
23