AI is creating far more questions than search ever handled. Google says queries are at an all-time high.
But each answer names only a handful of businesses. Rand Fishkin's research with SparkToro and Similarweb shows most Google searches now end without a click.
This week's stories show agents acting on those answers.
In one of them the agent books the dentist itself.
That makes the shortlist shorter still, and sometimes the agent picks one business and goes ahead.
via Elon Musk (@elonmusk), Noah Shinn (@noahrshinn) and @Texas_Twins_Dad
Three posts this week point the same way.
Elon told people to have their agent buy from sites other than Amazon. Noah Shinn's team at Instinct launched hotel booking done by agents. And one user said their agent found a nearby dentist that takes their insurance, then booked the visit.
That last one is a single post from a small account, and I want to see whether more people report the same thing.
EMElon Musk@elonmusk
Just ask Grok @Bot to buy what you want someplace else and get a better deal.
There are far more purchasing websites than just Amazon!
Instinct should be simple. You shouldn’t have to work around someone else’s app or website to get something done. The interface should be built for you, every time.
Today we’re rolling out a new hotel experience, built on exactly that idea. Tell Instinct where you’re going and …
My primary Grok Bot searched for dentists near me for my son, found one that accepts our insurance, and made the appointment on its own from its web browser for the first time. I got the confirmation email and text. Done.
I really don’t like calling dentists to check insurance …
After 48 hours with founders and lab people in San Francisco, Greg's top takeaway was "sooo many stories" of brokers selling company data to AI labs for around $2M a week.
That's his anecdote, not a sourced figure. He says himself he can't tell if it's the start of something or a bubble.
I don't need the exact number to believe the direction.
GIGREG ISENBERG@gregisenberg
I spent 48 hours in SF hanging out with billionaires, data labs and AI founders and execs running $1B to $100B companies.
My observations:
1. Sooo MANY stories of data brokers making like $2M a WEEK selling company data to labs. I can't tell if it's the beginning or a bubble, …
Patrick argues that companies have long used their "cognitive surplus" against shoppers, with confusing pricing and subscriptions that are hard to cancel.
A personal agent reads every line of the fine print. It doesn't get tired.
I think he's right, and it makes me hopeful for shoppers.
PCPatrick Collison@patrickc
Assorted things I've been mulling over regarding agents in the economy:
• Companies are in some sense the original superintelligence. It's often the case that they can use their amortized cognitive surplus to advance their interests against those of certain consumers: relative
by Pieter Levels (@levelsio), with a counterpoint from Lenny Rachitsky (@lennysan)
Pieter lists the moats he thinks survive cheap software, which are proprietary data or content, network effects, on-demand distribution and something physical.
Lenny pushes back that most so-called moats, distribution included, aren't real.
I'm with Pieter on one of them.
L@levelsio@levelsio
I'd expand the moats left to:
- Proprietary data or content: data nobody else has, or unique content you can create - Network effects / community: users make it more valuable for each other - Audience / distribution: you can get attention on demand - Physical component:
Shoppers can now connect any agent that supports MCP to their Shop account. It searches Shopify's catalog, checks out with Shop Pay and tracks orders. Every purchase needs approval by default, with an optional budget mode.
If I sold products online, I'd study this launch first.
The store now plugs straight into the shopper's own agent as a skill.
TLtobi lutke@tobi
Turn any existing Agent into the world's best personal shopper for you.
The new shop skill: https://t.co/2cyHQms4Vd
Add the skill & connect your Shop account & finally find the thing that you always wanted …
According to Aaron, most work today is limited by people typing prompts, and moving agents into the background could raise token use 100 to 1,000 times.
Agents running in the background would read your company's data around the clock.
ALAaron Levie@levie
Moving from chatting back and forth with agents, to having agents that you spin up in the background to do useful work for you is probably on the order of 100X or 1000X more token consumption than what we’re seeing today.
Most work right now is essentially rate limited by our
