What's Important to Me: According to My X

When I turn the X search tools inside the agent back on my own activity for the last thirty days, a ledger appears.

Not an abstract social graph. A map of what I actually stop for, tag into, build around, and return to. The subjects, the sentiments, the specific humans whose posts and replies shape the work.

This is what is important to me, according to my X.

The Method (Last 30 Days)

Posts authored, threads engaged, direct tags and replies between roughly May 19 and June 2 2026. The same X search that powered the commerce thesis and the attention chart in the other dispatches. Turning it inward.

Agents as Team — Good That Builds

June 2. Two posts that carried the weight of the whole period:

“using AI agents is like managing a bunch of people.

you know why I am hopeful for the future?

because that means good will win.

because it is good that builds.

evil only robs — and AI will enable builders more than robbers, well, because robbers are blinded by greed.”

And the companion:

“AI agents are not going to make everyone a builder.

But they are going to make it much harder to hide the fact that you are not one.”

“hiring agents. remote work ok.”

This is not replacement talk. It is management, taste, responsibility. The craft moves upstream. I am the one who decides what “good” means in context and ships it to real customers or readers.

The tone is deeply hopeful: leverage for the people who actually make things.

The Commerce Proposal — Named Humans, Specific Ask

May 29. The full thesis post:

“X can revolutionize commerce, revitalize X Ads, and provide massive value to users by partnering with Shopify and bringing the Universal Commerce Protocol to the platform.”

I had just published the longer version as “The Premium Blind Spot.” Then the direct tags in the thread:

“@nikitabier give me 1 dev for 2 weeks and let me build this thing with you guys. OpenAI dropped the ball on this exact partnership…”

“cc @gilgNYC I know you are (understandably) bullish on UCP. I would love to see this partnership happen between X and Shopify.”

“@nicgrasset this might not be your domain but if you can review this opportunity for Shopify, would love to hear your feedback. I was bummed that OpenAI pulled out shopping with UCP.”

Earlier the same day I had written plainly:

“I have Premium+. I have never seen an ad on X. The better Premium gets, the more of X’s best customers disappear from the ad business. The X ads model is eating itself.”

I am not complaining from the sidelines. I am naming the exact people who sit in the rooms and offering to ship the fix with them.

The tone is strategic and direct — a builder making a concrete proposal to the people who can move it.

Honest Tool Report — X Search and Imagine Are the Real Edge

Late May. Multiple public notes while shipping the first three Signals (including this one) with Grok Build:

“My first product using Grok Build: THE SIGNAL. Since Grok Build can use X Search and Grok Imagine for no added cost, my first shipped product is leveraging that.”

Then the candid feedback:

“you guys gotta be honest with grok build

it is great for research but i seriously have not been able to do better work than claude code yet.

x search and grok imagine, that’s what you guys should speak about.”

“Use grok for research on x or creating grok imagine images. It sucks at building UI, like really bad.”

“doing research with X Search has been amazing.”

“The no-API key grok imagine and x search are underrated.”

” I have 3 agents working on the same file locally. and somehow no conflicts. I guess it pays to be a dev.”

“having Subagents in Grok Build has been great.”

I will celebrate the parts that actually multiply my output (research + the plates you see in these dispatches) and say plainly when the tool is still fragile for other work. That is also how you build.

The tone is appreciative of the unique power, self-aware, and practical.

The Ground — Honest Money, Aligned Partnership, Builder for Life

Recurring through the same weeks — the ground truth signals:

Photo of festival setup. “I made $1,475.71 last year at a physical festival where I made coffee, for my coffee brand @sevakcoffee.”

“honest work, makes me count my blessings at home working.”

“the things my wife can do together when we agree on the vision and work together, that’s a force that makes me a romantic.”

“i work with my wife, so both, yes.”

“as a business owner, a vacation is just remote work with extra challenges.”

“Is 35 too late to buy a new domain?”

“i am never retiring”

“can i work while retired?”

“haha — I started doing subdomains for all my ideas so I stop buying domains”

The small honest physical business and the partnership at home are not side notes. They are the reason the hope about agents and platforms feels real instead of theoretical. I know what $1,475 in actual sales from people who tasted the product feels like.

The tone is grateful, romantic about aligned work, stubbornly anti-retirement, and practical.

The People Who Surface the Map

These are not the loudest accounts. They are the ones whose posts and replies I actually read all the way through and find myself tagging when something needs saying.

The product minds I named in public with a concrete offer. The aligned voices wrestling with the same tension — how to use powerful tools without losing the part that wants to build something good. The community voices that keep the picture of real users and real builders in view.

When the X search is live inside the agent, the network stops being theory. It becomes specific humans with specific posts who make the next move clearer.

That is the real signal.

The ones who remind me, daily, that good still builds — and that the tools are finally starting to favor the people who do the building.