websets what
hi jeff!
it’s been a while. last time we spoke, grok wasn’t even a thing lol.
things went pretty damn fast. ai is ubiquitous.
everyone is now obsessed with capturing the audience:
- grok has x
- perplexity builds their own browser
- claude has a cli tool
- gemini on every fucking rounded corner google touches…
it won’t be long before ms adds copilot in the bios.
open ai doesn’t really care because “chat gpt” is synonymous with ai for most people, and normies will gleefully search for it on fucking bing to do the funny ghibli picture like all of their friends.
it seems that distribution does matter.
and exa? good old-fashioned emailing campaigns!
this sparks joy. let’s see that:
your email
monitors is dead on arrival. sorry.
the ripple effect is now:
original happening -> x -> google -> others (including news)
and there’s no beating grok with ifttt hacks. maybe it’s nice if you’re already a websets user though, i wouldn’t know.
data export, i can understand, but data import…? to upload some data and refine it, i guess. not bad! could be a very efficient gateway!
the words “csv” and “salesforce” scream corpo. we’re far from the hardcore devs fan base.
full api? email coverage? what? obscure.
exclude existing data… i understand the feature but shouldn’t this be baseline?
websets use is already “pay per result”, this conveys the idea that you’ll pay more than you should.
the promo: -30% on a thing of unknown price is not sexy.
i know some stuff: the price of bread, rent, nintendo switch2, and that “premium ai” is 19.99$/month.
i think 20 bucks the first month would be better messaging: “oh, so this thing is normally more value than gpt plus? interesting.”
my two cents. let’s go to the website to check the new prices.
websets website
first, websets and exa is a muddy separation. search isn’t the flagship product anymore?
the first examples are really, really great, and i find them highly converting for carol.
but then again, carol is probably corpo now (with a big sales navigator budget) so “research” is not on target.
enrichment: i like that! builds on the “supercharge query” that i sent you. valuable.
verified: not sure about the wording… carol cares if the emails she gets are genuine, not about “agents”. people know that llm outputs cannot be taken at face value, you have to give them more than that.
api is now at the bottom, sad but coherent with the strategy.
what now
well, now i do youtube videos!
building fun full-stack projects with python.
i am trying to solve delivery, and throwing ideas and projects out there to see what sticks.
i’d gladly make one about exa if you want ;) just tell me.
i don’t know if the strategy pivot was a good idea, but i can tell it was greatly executed. well done.
i hope this works out!
cheers,
le