Hey guys, I’ve been using Apple’s Voice Memos app to capture ideas on the go, for quick thoughts, content ideas, software ideas, random “I should build this” moments. The problem: Apple's Voice Memos app is just a recorder. My notes stayed trapped as audio files. No organization, no summaries, no action items, no way to actually use what I captured. So I built Echo. What Echo is: A personal voice memo system with an AI layer on top. You record a voice note right in the app, and it: - Transcribes it - Summarizes the key points - Pulls out action items - Tags and categorizes it (content ideas, notes, reminders, etc.) - Stores everything in one searchable inbox Since everything gets logged into my Supabase, i have full control of what I can do with my data. And, I can use both the webapp and IOS app versions whenever I need to! Disclaimer: Echo is not on the app store (yet haha) i'm using the dev build from XCode. If you’re someone who thinks out loud and wishes your voice notes were smarter, this is exactly the problem I was solving.

Yooooo Michael T. just spilled the beans on Cursor's newest Composer model and its going to be amazing. Here are the 3 things he says about their newest model: 1. Massive Scale and Compute Size The next model scales up engineering resources immensely to compete directly with the world's largest flagship models. - 1.5T+ Parameters: As shown on the presentation slide, the model features well over 1.5 trillion parameters, putting it on par with OpenAI's massive scale. - 10x to 20x Compute Power: They are leveraging significantly more compute than ever before. Past architectures (like Composer 1 and 2.5) were severely bottle-necked by running on minimal GPU clusters for shorter periods. This massive computing upgrade aims to hit and surpass the current AI frontier. 2. Trained Entirely from Scratch The model abandons third-party open-source baselines to build a custom foundation from the ground up. - 100k+ GPUs: The model is being pre-trained on a massive network of over 100,000 GPUs. - Granular Behavioral Control: While expressing appreciation for open source, the speaker emphasizes that building from scratch grants them complete control over model behaviors. This allows them to cleanly optimize the ecosystem specifically for developer workloads. 3. General Intelligence Beyond Coding The model is evolving past basic code generation to function as an autonomous, full-fledged software engineer colleague. - Expanding the Toolbox: The model handles tasks past standard syntax writing, including long-term planning and interacting directly with developer tools. - Actionable Execution & Testing: The AI will actively test software, click interface buttons to debug, and utilize an advanced user experience (UX) to explicitly show users everything changed by the agent. Guys, this is going to be big! This along with their new Github competitor: Origin ANNNND their new iOS app, I'm so stoked to see all those Claude Code users all of a sudden shift over to Cursor... not like i've been saying this for a year now.

@Nate Schweitzer sent me this gem today! Same task. 3 different engines. Only one winner And that winner is cursor. Seriously guys if you are building your agency, your automations, your whatever Cursor does it better faster and cheaper Their composer 2.5 is AMAZING. The entire Ghostsuite has been built purely with composer and I’d never go back. So far I’ve converted the likes of @Jackson Oaks , @Nate Schweitzer and @Doug Grosch to the holy land I’m bot alone anymore People are waking up from the matrix You next? Come check out cursor and get the first month free: https://cursor.com/referral?code=AFABQWJMKIBM

Brooo!!! Literally cooking rn on Ghostrack, the newest edition to the Ghostsuite family. As Ghostcoded is expanding and growing, I've found myself getting stretched thin on the communicating progress of projects And I don't like anything on the market bc... it's all old and not ai first... And I didn't build it so it's not how *|* want and need it to be. So I went ahead and built one myself. Some features include: - building client projects to organize everything - adding phases, sub phases, and checkpoints that you can mark as completed etc - publish updates that will appear in the clients very own customer portal - internal tasks - needs from client - automatic email notifications for progress updates - oh and **whats this**: an Al PM feature to help your ADHD brain focus and stay on task I couldn't help getting so many things twisted in my brain and so I solved my own problem, And maybe it'll solve yours too Go play with the beta free track.ghostcoded.com The Ghostsuite didn't start bc I wanted to just sell random software to agencies Each one was a massive pain point that I not only felt but was bleeding from I had to fix it and so I just bit the bullet and built it all for myself And I haven't stopped using them all since So if no one else wants to use them, I'm all good bc it's helping me improve my own business! Feel the pain before automating it

What's up everyone! One of the most powerful things you can know about when building ai automations, agents and software is: APIs. I'd venture to say that 95% of people who are "building" ai don't know what APIs are, how they work, or what they do. They might argue that Claude will handle that for them, or that Cursor just knows what to do. News flash: They don't. and from personal experience, a lot of times, they lie when you ask if you can connect with a certain app in a certain way. This "certain way" is the protocol known as an API (Application Programmable Interface). This protocol runs the internet. when you visit a website, API. When you pay your bills online, API. When you open LinkedIn, APIs. this protocol is the backbone of majority of communication on the internet and no one even knows what it is or how it works.. its just magic!! "But tanner, like aren't they right? Do we really need to know about apis?" Yes. When you are working on a consulting project, the first question is going to be: "Can we connect with my system" Your first thing to look up is: "Does x system have an api". Then from here you are able to see if their platform even has an api, is it a paid one, and how complex or easy is it to work with. sure an ai can do this research for you, but word of caution: They want to tell you what you want to hear. I have been lied to soooo many times by ai saying that a particular endpoint exists that does "exactly what i want it to do" when a matter of fact, it doesn't. so when your building or consulting, apis are very relevant and powerful to understand. So for these reasons, I've done some videos on breaking down apis and using n8n as the medium. Any questions? (LMK if you like this kind of content and if it is helpful or not)