Pinned Tweet
fuck it, I'm starting my own hackathon. Owned by the builders who enter it. And (unlike @marclou ), I'm putting up my own money ($2,500) to start the prize pool.
free to enter. I never touch the money. and the builders vote on the winners in public.
Higgsfield gave me $100,000 to realize my dream:
A hackathon where participants build a real business, not just an app.
So... Introducing The $100K Hackathon:
🏴☠️ SHIP-OR-DIE.COM/100K-HACKATH… 🏴☠️
It starts on November 16, and you'll have 30 days to build a startup and make it profitable. The prizes:
🥇 $100,000
🥈 iPhone Duo
🥉 MacBook Neo
This is not a joke. The winner will get $100,000 in cash.
To win, you must:
1. Start a completely new business on Nov 16
2. Build a digital product (SaaS, directory, etc.)
3. Make the most profit (not revenue)
You can already register at $99. When the hackathon starts on Nov 16, you can still join for $149.
I'm running this with my dear friend @jackfriks, and we'll manually review each potential winner to make sure everything holds up.
See you on November 16, builders 🏴☠️🫡
Readers added context they thought people might want to know
Under X’s Paid Partnerships Policy, organic posts that are part of a paid partnership must have the official “Paid Partnership” disclosure toggle enabled in Content Disclosure settings. This post does not appear to include that required label.
help.x.com/en/rules-and-p…
This is exactly my motivation for antipaywall.org
Here's my thought on the future of software/SaaS.
I'm completely switching off hosted platforms and hosted tools. I just reverse engineer them for myself, because I only need parts of them.
Two reasons why:
1. I want to own my data.
I use my data in different tools. Marketing, note-taking, whatever. And I want my agents, like Claude, to have complete access to it.
Not walled off in HubSpot, where I have to talk to their agent. Not behind MCP either. With MCP the agent pulls what it asks for, but it doesn't know what's there.
Give it a full list of data files and it has complete access. It knows what's there.
2. I want to customize it.
When I sign up for Pipedrive, I don't need all that bloated shit. I need one part of it, adjusted to my content, my business, how I work.
With AI that's a prompt, and one hour later it's done.
So static software is going to be a thing of the past. People want to adjust it to how they work and plug it into the other tools they have.
Where this goes, I think two ways:
A. You buy software again. Like paying $900 for a Photoshop license and putting it on your computer. Except now you get the source code. It's premade, because you don't have weeks to build it, but you host it on your own headless box and you can customize it, remove stuff, plug stuff in. Someone needs to build the operating system for this.
B. Everyone gets their own hosted instance. Not 10,000 customers on one box. Your own sandbox, with an agent that customizes it fully to you. The code stays proprietary, but you can adjust everything.
Might be completely wrong. Might be different for most people.
But this is already how I work.
32
Ethan+ retweeted
Here's my thought on the future of software/SaaS.
I'm completely switching off hosted platforms and hosted tools. I just reverse engineer them for myself, because I only need parts of them.
Two reasons why:
1. I want to own my data.
I use my data in different tools. Marketing, note-taking, whatever. And I want my agents, like Claude, to have complete access to it.
Not walled off in HubSpot, where I have to talk to their agent. Not behind MCP either. With MCP the agent pulls what it asks for, but it doesn't know what's there.
Give it a full list of data files and it has complete access. It knows what's there.
2. I want to customize it.
When I sign up for Pipedrive, I don't need all that bloated shit. I need one part of it, adjusted to my content, my business, how I work.
With AI that's a prompt, and one hour later it's done.
So static software is going to be a thing of the past. People want to adjust it to how they work and plug it into the other tools they have.
Where this goes, I think two ways:
A. You buy software again. Like paying $900 for a Photoshop license and putting it on your computer. Except now you get the source code. It's premade, because you don't have weeks to build it, but you host it on your own headless box and you can customize it, remove stuff, plug stuff in. Someone needs to build the operating system for this.
B. Everyone gets their own hosted instance. Not 10,000 customers on one box. Your own sandbox, with an agent that customizes it fully to you. The code stays proprietary, but you can adjust everything.
Might be completely wrong. Might be different for most people.
But this is already how I work.
I’ve been using @crowdhealth for years. It’s how health insurance should work. I pay $50 a month and have an option to contribute to someone else’s coverage once a month. I can say no whenever I want.
Personally I haven’t had to use it. But a friend that joined after my recommendation was suddenly hospitalized after just a few weeks of joining. His hospital bill was $180k. Crowd health funded it no problem. All he paid was $550 ($50 for first month + the first $500)
$50,000 bill
You pay the first $500
We ask the community for the remaining $49,500
495 people will be asked for $100 (for example). We only ask members once per month to help another member*.
That $100 (x 495) will go from their account to the account you set up during the sign up process.
If someone’s payment method fails or declines to fund your bill, we ask the next person on the list to fill it.
One you have the $49,500 in your account (takes about 7 days) you use that to pay your bill.
68,000 times and counting. 99.9% success rate getting these bills funded.
*if your payment method fails or you decline a bill we will give you another opportunity to fund a bill.
2
65
The real product has grown to a full sales suite for agents. Voice really brings it all together. Here is a preview. Full product reveal coming the next few days.
I launched the @sellbirdagent brand & website 14 days ago.
- 12,950 website visitors
- 1,384 early access requests
- 20,000+ brand searches
- 6.9k stars on Github
This was all organic. I've done zero marketing myself.
The real “marketing hack” is just having a good product.
1
2
192
I launched the @sellbirdagent brand & website 14 days ago.
- 12,950 website visitors
- 1,384 early access requests
- 20,000+ brand searches
- 6.9k stars on Github
This was all organic. I've done zero marketing myself.
The real “marketing hack” is just having a good product.
Good update @marclou
Two big updates to the $100K Hackathon 🏴☠️
1/ We have disqualified the #1 product.
After reviewing the entry, we found that work started before November 16 and that it was a reworked version of an existing product.
Both break the requirement to build a completely new business from zero after the hackathon starts.
This decision applies to the product. Jacky can still participate with a new business that meets the rules. The disqualified product’s revenue won’t carry over.
Here's the updated leaderboard: ship-or-die.com/100k-hackath…
2/ We’re changing how the $100,000 is distributed.
I’ve been reading your feedback. Having one person win all the cash made the prize feel out of reach for many of you.
My goal with this hackathon is to get people to finally ship the business they’ve been thinking about. If more chances to win help more people keep going, that makes sense to me.
So we’re replacing the single $100K cash prize with:
🥇 1st: $30,000
🥈 2nd: $20,000
🥉 3rd: $10,000
4th: $8,000
5th: $7,000
6th: $5,000
7th: $5,000
8th: $5,000
9th: $5,000
10th: $5,000
Plus the iPhone Duo and MacBook Neo. 12 winners in total!
The cash pool stays at $100,000. Winners are still ranked by verified profit.
Building apps changed my life. I’m doing this with @jackfriks because I want more people to experience that too.
Thank you for helping us make this hackathon better 🫡
84
yikes
Replying to @cb_doge
?
We aren’t serving open source models
Readers added context they thought people might want to know
Cursor serves hosted open-weight models from Chinese providers including GLM 5.2 (documented as served via Fireworks) and has announced GLM 5.3/Flash and DeepSeek V4.1 Flash availability; it is only exploring phasing them out.
cursor.com/docs/models/gl…
x.com/cursor_ai/stat…
x.com/cursor_ai/stat…
cnbc.com/2026/11/20/cur…
67
If you run a service business and aren’t making way more money with AI you are doing it wrong.
My lead to close rate is nearly 100% for services I sell.
this is how:
1. Productize your services. You need to know exactly what your services and pricing are with SOPs connected to each one. If you can’t sell it like SaaS then you don’t understand it enough.
2. Only get on calls with people you know will close, which is qualified by your intake system.
3. Stop using forms. No more “Get a strategy” or “Request a call” CTAs. That’s basic shit. you’re better than that. you need to let them experience your expertise through an interactive intake. "Start now" next to a domain field is the CTA that starts the intake.
4. The intake is mostly weeding out who you DONT want to sell to. Every answer collects more data and routes them down a unique path. If you think "but I can sell this to anyone" you're ngmi
5. For people who are not qualified, their info gets submitted and they see “Doesn’t seem like a good fit" but reach out to us here if you think we got it wrong.
6. qualified people get to book a call at the end. “but what if they don’t complete the intake!” then they aren’t a good fit, and you have their info already anyway if you want to still sell.
7. your sales calls should be a demo. Not a generic demo, the demo of their service. If it’s a website, show them the website. if it’s marketing show them your strategy and ad creatives. This is made possible through agents with defined skills kicked off through steps in your interactive intake system. it doesn’t have to be perfect, you will tell them this is what we do BEFORE we work together. they will wonder what the paid version looks like and ask to buy.
8. send the proposal/payment on the call. when you show that you (your agents) have already done most of the work they will buy. no one else is doing this. so show up prepared and when they are bought in the payment link is already in their inbox.
9. All of this is made possible by having a single OS with MCP. you connect all your tools and agents to it so everything is connected and you have one source to connect to your agent, not 20, and its all connected. This makes you move at the speed of request.
10. deliver good work. the job is not done when they pay. the job is never done. your role is to provide a valuable service month after month so they continue to pay and become an advocate of your service that sends you referrals.
Here is the visual I showed for this if you want to explore: claude.ai/artifact/EgiVLMPbC…
69
how is that OpenAI released Astra and it was CRAZY good for like 3 days and then just completely down hill since
1
149
Ethan+ retweeted
Interesting trend here.
Startup founders at major SaaS companies are desperately trying to build an audience/brand.
Why? Software is now musical chairs. We all know the music is about to stop, and nobody wants to be caught without an audience.
I think what @indexsy has done is fine. It's lame imo, but he's not doing anything objectively wrong. He saw an opportunity to grow a following + make money and was able to do so because he already built an audience. Incredible.
The problem is the embarrassing lack of planning by @marclou and @jackfriks that enabled @indexsy to easily be in first place. Imagine trying to charge your followers $99 to participate in a contest they have no chance of winning...
1. They tried to sell access to this "hackathon" without having any rules in place to keep the competition fair. Both of them should know that with their audience (who they interact with every day) this would easily lead to hundreds of less experienced people rushing to buy whatever they put in front of them. None of these people ever stood a chance to win anything.
2. selling access to a competition with a prize pool is literally gambling. It's against Stripe ToS to even process those payments.
3. After pushback, Marc posts a "Heard your feedback, you're right, I refunded everyone!" spun as an innocent oversight.
4. Marc has several businesses on Stripe with his most successful one, TrustMRR, relying completely on it. Many people including myself pointed out that those payments were in breach of @stripe's terms. Stripe bans accounts all the time for doing the exact same thing. Marc even has a personal relationship with Stripe. They've apparently flown him out to events and publicly congratulated him. So between Marc not wanting to ruin his relationship with them, and them having a direct line of communication with him already, what are the odds he refunded people from the goodness of his heart vs being forced to refund everyone...?
5. The entire thing is a marketing funnel for their paid course / membership Ship or Die. That part is fine (lame af but get your bag i guess). The problem is that when this was launched and you had to pay to get into the contents, existing Ship or Die members GOT FREE ACCESS TO THE CONTEST.
6. Jack and Marc have huge audiences and profitable products. They host a contest with a paid prize, and not only do they not use any of their own money for the prize pool (a "PRIZE" for THEIR audience), they get the prize money from the most dishonest company in the build space, Higgsfield.
7. Even calling this a hackathon is dishonest. This is a cash grab contest. And it seems intentionally designed to suck as much cash out of the followers that enabled all of these "influencers" to be where they are today.
But I don't want to assume. "Never attribute to malice to what can better be explained by incompetence".
So @marclou @jackfriks which is it?
woke up to a ton of hate, hilarious, some notes:
- where tf are all the indiehacker or seo influencers who talk a big game?
- a lot of you need to fix your mindset, you ngmi with that one bro
- i carved out a villain persona because i knew how sensitive indiehackers are (complaining about $99 fee lol)
- ty for virality, your hate is only making me stronger, seems like i dont even need to run ads anymore
- i didn't cheat, i did a lot of planning in the days leading up to it, and i knew my github will be heavily audited lol
- no i didn't ask my friends to buy it, i just used my audience + newsletter, not against the rules
- no i didn't use any os repos or existing products, prompt was use shadcn + dataforseo and build a local rank tracker lol
- i'd still beat you all without an audience because i know how to run ads lol, i have a 26k mrr saas in stealth
Ethan+ retweeted
We’re roughly 17 years into crypto usage and the only widespread use cases are criminal activity, scams, and as a gambling instrument.
I’m conflicted on this. I hate regulation and love AI but food advertising seems like something that should not be generated content
I made a 27 second AI UGC ad for a Texas burrito chain for about $10 in credits, and Claude Code ran the whole thing from one terminal
Restaurant UGC is the most formulaic ad format there is. Burrito next to the face, big first bite, queso pour, city names at the end. Which makes it perfect for agents
Here’s the whole workflow:
1. Pulled their live Meta ads with the ScrapeCreators Ad Library endpoint through Graphed Tools. They had 1 ad running, a holiday gift card promo. No food in it at all
2. Scraped their last 48 Instagram posts (77 images) with ScrapeCreators. These became the food references, so the AI food is their actual steak burrito in the red tortilla and their actual fish tacos, not a stock burrito
3. Wrote the script. Hook on camera, voiceover over food b-roll, CTA on camera. 66 words, 9 shots
4. GPT Image 2 put our avatar Maria in a booth in their store, with the surf mural and red stools from their IG, holding the burrito. Then 5 food start frames and an end card. 8 images, about 30 credits
5. Her cloned voice read the script with fal MiniMax TTS
6. Seedance 2.5 rendered the talking clips at 480p from the still + the MP3. She says the hook then takes a real bite. Then 5 food clips at 720p from the start frames: queso pour, burrito cut in half, lime on the fish tacos, chip dunked in queso
7. whisperx checked every word Seedance said against the script. 0 wrong words ffmpeg cut it on the voiceover word timings, burned in word by word captions, music under it, end card
About $9.30 credits for the video, most of it the 720p food clips
Every vendor API ran through the Graphed CLI on one credit balance. No fal key, no kie key, no ScrapeCreators account, graphed:transcribe
PRO TIP:
scraping their IG first is the whole trick. The food on screen is their food, so it doesn’t read as AI slop
you can make an ad just like this
just get started for free here - graphed.com/
and give you claude code this post text
2
4
399






















