Michael Ho

co-founder @D3inc, scaling the @domaprotocol 🌐. DomainFi for the internet’s 360M+ domains. ex-@goldmansachs

[Michael Ho @mxchaelho]

@solana, meet domains 📷

Tokenized exotics are REAL, and domains are the first in this Supercycle via @domaprotocol.

This was a fun one, originated by our community. Very fitting for the trading audience.

More REAL assets coming on-chain very soon...

[Michael Ho @mxchaelho]

domains are coming to @solana.

domains enable internet capital markets, quite literally.

[Sunrise @sunrise]

The internet runs on names. Domain names are assets.

Sunrise is partnering with @DomaProtocol by @D3inc to support the arrival of tokenized domain assets on Solana.

A new category of RWAs is coming onchain.

[Michael Ho @mxchaelho]

a great example of the status of the web.

most sites are still NOT equipped to handle agents.

this does not improve with the ever increasing amount of agent apps (everyone uses different standards).

[CrowdReply @CrowdReply_io]

Ramp, a $60B company, hid a secret offer on its website that only AI could see.

in 32 days, AI told buyers about it 370 times.

here's the genius AEO experiment behind it:

1. They built a second version of their site just for AI

- Ramp put an offer on about 50 pages that only AI bots could see
- Human visitors saw nothing different
- Then they tracked whether ChatGPT, Claude and Perplexity would pass the offer on to real buyers

2. They tested 3 formats to see what AI actually reads

- Markdown
- Plain stripped down HTML
- Pages packed with schema, the code made for machines
- Everyone expected schema to win
- Markdown was the only format that reliably showed up in AI answers

3. They found out who was really visiting their site

- Thousands of visits from 7 different AI platforms in just 2 weeks
- Some AI bots pretend to be normal browsers. DeepSeek shows up as a Chrome version from 2017
- Claude's crawler visited more than all the other named AI crawlers combined

4. They asked the AI tools buyer questions every single day

- The same questions a CFO would ask when picking a spend management tool
- Then they checked every answer for their offer

5. Every AI behaved completely differently

- Claude passed on the exact offer, the link and step-by-step instructions
- Perplexity only mentioned it vaguely
- ChatGPT read the pages but mentioned the offer 0 times in 32 days
- One strategy for "AI" doesn't work. Each model needs its own plan

6. The biggest lesson: AI trusts pages it already cites

- Pages that AI was already citing were the ones where the offer showed up
- Pages with few AI citations got zero mentions, whatever the format
- One of the best pages was one Ramp never cared about for Google

the result:

- 1,300+ AI bot visits in the first 2.5 weeks
- about 40 AI answers mentioning the offer at 2.5 weeks
- 370 by day 32, almost 10x
- a peak of 33 in a single day

Ramp didn't win this by writing more blogs

they figured out which pages AI already trusts, then put the right message there

and that's the whole game now: if AI doesn't already trust the pages talking about you, it won't recommend you.

if you want to know which pages ChatGPT and Claude trust in your niche, and get your brand mentioned on them.

reply "citations"

[Michael Ho @mxchaelho]

deck making on @claudeai artifacts is like talking to the perfect analyst we never had… at 100x the speed.

days of work turn into minutes, literally 🤯

good tech.

[Michael Ho @mxchaelho]

@cultra @godaddyauctions That’s a good deal!

[Michael Ho @mxchaelho]

dns + agents is still massively broken

we are upgrading the internet for what’s to come @D3inc

any domain, any agent welcome at https://t.co/NdrzGXJsIO

[Maia in Palm Springs @maiab]

re: agentic shopping
I have NEVER in my life heard so much hype and discussion about a tech/product experience that basically does not work at all for me.
I feel like I’m using theranos

[Michael Ho @mxchaelho]

Domains aren't just websites. They're endpoints and rails for humans and agents.

Agents are driving the next wave of internet commerce.

@domaprotocol brings liquidity, and DNA brings interoperability for agents on the Internet.

[D3 @D3inc]

Your customers stopped browsing. They send @Muse, ChatGPT and other agents instead.

DNA lets you see every agent on your site and decide who gets in and who pays.

Earn from AI → https://t.co/B3J7poIbgo

[Michael Ho @mxchaelho]

A new SOL ☀️ is coming on the DNS 🌐

[Doma Protocol @domaprotocol]

doma.sol

[Michael Ho @mxchaelho]

The domain SUPERCYCLE is here.

[Doma Protocol @domaprotocol]

Land rush. Domain rush. Domain supercycle.

Don’t watch this one from the sidelines 👇
https://t.co/tiWZ7yxLtI

[Michael Ho @mxchaelho]

Joining @stabledash Live today at 2:30 PM ET to talk about Frontier and new internet domains with @D3inc!

https://t.co/wYO5vx9F7s

[Michael Ho @mxchaelho]

@pitdesi Cote at Palazzo is on par+ with NYC.

[Michael Ho @mxchaelho]

@seyong @solquicks

[Michael Ho @mxchaelho]

@chelms34 congrats!!

[Michael Ho @mxchaelho]

@solquicks 🥺

[Michael Ho @mxchaelho]

👀

[D3 @D3inc]

Hint: a whole new ecosystem is about to join the next internet land rush.

[Michael Ho @mxchaelho]

Can’t imagine what the line looks like if they added RH’s prediction markets biz here.

Likely 2-4x the other apps combined. RH controls the distribution and is increasingly capturing half of the exchange economics too (moat).

As a shareholder, love it. As a user… 🫠

[Quant Chad @Autonomous_Chad]

Retail lost $280M trading on Hyperliquid

The 400k wallets trading through 3rd party apps (Phantom, Metamask etc.) are down -$712 each on average.

This is the reason Hyperliquid is winning so hard

Let me explain

1. RETAIL FLOW ATTRACTS WHALES

Whales need dumb counter parties.

Trading against other whales is hard and risky. No one wants that.

To be consistently profitable, whales want unsophisticated retails making bad bets to trade against.

This is called “soft flow”

2. SOFT FLOW CREATES REVENUE

-> More soft flow
-> more food for whales
-> more trading volume
-> more fees for Hyperliquid.

PerpDex and Prediction Markets will tell you that, contrary to Sports book, they don't trade against their user.

It's true in absolute but it doesn't mean that they don't profit from their losses.

If there was no losers (retail), whales would eat each other, become unprofitable and ultimately leave the platform.

Retail losses are the blood flow of any trading platform.

The more money retail loses, the more whales will come to trade and more fees the platform will collect so the whole game is about attracting more losers possible to your platform.

3. HYPERLIQUID HAS THE BEST MODEL FOR ACQUIRING RETAIL

HIP-3 and now HIP-4 are build on one founding principle.

Permisionless deployment. I.E letting independent builders create markets.

Each builder can target a specific retail audience :
- sports fans
- regional communities
- crypto traders
- finance bros
etc.

That's a huge distribution advantage.

Whales move between market categories quickly as long as there is retail to trade against, but acquiring retail audience for a new market category is what takes time and work.

HIP-4 let independent deployers do that specialized work for their target segment while the platform collects the benefits.

That's why they won perps. That's how they will win Prediction Markets

[Michael Ho @mxchaelho]

push, pull, or something new?

a new era for monetization on the internet is imminent. it’s just fragmented at the moment.

[Fred Hsu @heyheyfred]

You funded the AI agent - great. How does anyone know it’s yours?

25+ years in domains taught me this: a name opens doors. Trust keeps them open. We’re building D3 Edge to help websites check who’s knocking. Doma brings domain ownership onchain so those names work across the web, wallets and apps.

VCs, Web3 founders, wallets, chains: your agents need to get through someone else’s front door. That’s what we’re working on at @D3inc.

Who’s building with us? DM me

[Uncle BigBay @unclebigbay143]

@Porsche Okay, can it do this? 😬

[Michael Ho @mxchaelho]

@Xeer @Koreanguy_ @Aster_DEX @coinboy717 @ONEMINNFT 🫡🤌