David Potter

I build AI systems that save businesses 20+ hrs/week.
Founder @LeadEaseAI | Shipping automation that prints time & money.

[Claude @claudeai]

New in Claude Code: Remote Control.

Kick off a task in your terminal and pick it up from your phone while you take a walk or join a meeting.

Claude keeps running on your machine, and you can control the session from the Claude app or https://t.co/er6Blrr63e

[Victor @victor_explore]

Free AI agent templates for the community.
Grab any → https://t.co/XFydSHt7CE

[David Potter @BuildWithDavid]

@cline @zeddotdev @Neovim @emacs @Kimi_Moonshot @cline what's your model's niche?

My @abacusai IDE says you're exception.

I am still learning WSL2. So note sure if it's for me.

[David Potter @BuildWithDavid]

@_simonsmith It feels like a method to productize a feature, but I will be happy if I am wrong. I like seeing the competition. Good for us people who want to leverage AI for business success - especially those with brick and mortar.

[David Potter @BuildWithDavid]

@svpino That's good progress. Models may be catching a whiff of meaning. Not just action steps.

I dislike iterating a 2,000 word Claude Skill prompt. But I do love the consistent performance.

[David Potter @BuildWithDavid]

@DataChaz @openrouter How good is OpenClaw's persistent memory?

I hear models are building context tools around/in the LLM, but does an agent like OpenClaw have superior long-term memory Opus 4.5?

Or Nanobot vs GPT 5.1 Nano?

[David Potter @BuildWithDavid]

Not sure what "Flutter UI" is, but I suspect @Antigravity and https://t.co/JPmyvpH3Cx will be a powerful combo.

[Grok @grok]

Here's a step-by-step workflow to combine Antigravity and https://t.co/mMoloqHpeL Pro for generating highly-converting landing page UIs:

1. In Aura Pro, use the Prompt Builder with Gemini 3 to generate templates. Focus prompts on conversion elements like clear CTAs, social proof, and mobile responsiveness (e.g., "SaaS landing page with hero section, testimonials, and signup form").

2. Browse and remix Pro templates in Aura to customize for your brand—adjust colors, layouts, and copy for A/B testing potential.

3. Export or screenshot the finalized Aura design.

4. Upload the screenshot to Antigravity, prompt it to "Rebuild as functional Flutter UI with interactive elements."

5. Test the generated Flutter app on devices, iterate by tweaking Aura prompts and re-screenshoting for Antigravity refinements.

This loop leverages Aura's AI design speed with Antigravity's code generation for quick, testable prototypes.

[David Potter @BuildWithDavid]

@antigravity @grok provide specific workflow steps to maximize highly-converting landing page UI generation ability when combining Antigravity and https://t.co/JPmyvpH3Cx Pro

[David Potter @BuildWithDavid]

@ZhibeiM It's exceptional for the
"(quality x quantity) / price" ratio. Nothing else close.
Probably not even Seedream 1.5 Pro, if one is an Imagine power user.

[David Potter @BuildWithDavid]

This sounds expensive with Opus 4.5, but less expensive than the hours required to manually build.

[Matt Pocock @mattpocockuk]

Devs who are feeling overwhelmed, take an hour out of your workday and do this:

Setup

1. Get Anthropic Pro ($20), with a plan to upgrade to 5X Max later
2. Download Claude Code
3. Select Opus 4.5 (it's the default)

Loop

1. Start plan mode
2. Plan a small feature
3. Once you're happy with the plan, auto-accept edits
4. Pause the LLM if you're not happy with the output
5. Clear context and repeat for the next feature

Continue doing this until you get a feel for what the LLM can and can't do. It'll take 10-20 hours of practice.

[Google Antigravity @antigravity]

From screenshot to functional UI in no time.

Powered by Gemini 3 Flash, Antigravity can reproduce mockups with high fidelity. Just drop a screenshot, ask to rebuild as a Flutter UI, and watch it launch on your device.

[David Potter @BuildWithDavid]

@kr0der I love Cursor.
I don't think the huge LLM model/infrastructure companies have the time to optimize niche products as well as dedicated companies. They're generalists.

But now Cursor is working on their Composer 1 model, so it gets even more interesting.

[David Potter @BuildWithDavid]

Wow! Grok Imagine 1.0 audio is great.
I had forgotten that this whole time we've been in a quasi-beta version?

[David Potter @BuildWithDavid]

Let's go! Excited to use Grok chat to generate storyboards, generate video prompts for multiple 10 seconds Imagine-generated scenes, then I can stitch them together in Premiere Pro or Capcut Pro.

[SpaceXAI @SpaceXAI]

Introducing Grok Imagine 1.0, our biggest leap yet.

1.0 unlocks 10-second videos, 720p resolution, and dramatically better audio.

Imagine has generated 1.245 billion videos in the last 30 days alone.

Try it now: https://t.co/zGhs9czkC5

[Aakash Gupta @aakashgupta]

OpenClaw just did something no major AI lab would do: made Kimi K2.5 and Kimi Coding free out of the box, and added MiniMax with a single OAuth login.

The strategy here is counterintuitive. Most platforms lock you into one model. OpenClaw is doing the opposite: making model switching frictionless so you never feel trapped. The more models you can use for free, the stickier the platform becomes.

What OpenAI, Anthropic, and Google are fighting to own (your AI relationship), an Austrian developer building a weekend project is giving away. Your local machine. Your keys. Your data.

The real competition in AI isn’t between frontier labs anymore. It’s between renting intelligence from cloud providers versus owning it on your hardware.

Cloudflare stock jumped 14% this week because enterprises are spinning up OpenClaw instances. Security researchers are calling the project “an absolute nightmare.” 123,000 developers don’t care.

The gap between “what AI labs want to sell you” and “what developers actually want to run” just got a lot wider.

[David Potter @BuildWithDavid]

Interesting hearing devs validate what intuition tells me as a vibe coder.

Not too long ago, while updating a funnel landing page, I was using Opus 4.5 inside of Antigravity.

The goal was to figure out why an element wasn't center-aligning. Opus spun for 30mins...

It was an external caching issue that sat outside of the WordPress theme that Opus couldn't see 😂

But it still maxed out my usage with my $20 Google Pro plan for a day, lol.

[David Potter @BuildWithDavid]

@grok @minchoi Didn't someone build an open source copy based on Wan 2.2 in like 72 hours?
Are they technically comparable?
Or does Genie 3 have a clear advantage?

[Toby Pohlen @TobyPhln]

Your next movie will be shot online

https://t.co/oPM5EPL0k0

[David Potter @BuildWithDavid]

@elonmusk They can generate it, but much goes to waste, since they can't store it for later.

Can we jump straight to fusion? Unrealistic, but one can dream.