we just built our own help center for @brandjetai 🔥
used @ferndeskhq for a few months, genuinely loved the product. but we needed to push articles straight from our own admin dashboard
so we built one internally in 2 days using ai. full control over articles, publishing, everything from one place
these kinda things used to take weeks or months, but now with the help of AI it's just matter of hours or few days, crazy times...
I've put more into Artemis than anything I've built before.
People ask what I've been working on. This.
Artemis : the world's first AI agent that runs a full sales cycle from one prompt.
This is what sales infrastructure should have looked like all along, and I'm happy to finally be able to share this.
Introducing Artemis.
The worlds first AI agent that runs your entire GTM motion from a single prompt.
If you're running outreach, Artemis can make you $100k+ in 60 days.
Comment “ARTEMIS” and I’ll send you access for free.
@dramaricic for brandjet i just went where people already complained about the problem. reddit threads, twitter rants, 1 star reviews of competitors. then replied genuinely without pitching anything. took months but way better than cold email spam
@apoorv_taneja gpt-5.4 is interesting for the speed stuff but claude still resolves more real issues on swe-bench. curious if the api errors are anthropic side or just rate limits from heavy usage
been using claude for brandjet's backend stuff and honestly the api lockdown in january hit harder than expected. had to rework my whole flow. but the model quality is still better for my use case than chatgpt so i just pay the api tax and move on
mainstream consumer stuff was never their play anyway
@DmytroKrasun you don't need a different worldview. the metric is just disconnected from what you actually care about. some weeks i ship 500 lines, some weeks i delete 300 and the product gets better
@Bencera this assumes ai can actually own outcomes and not just execute tasks. tried giving ai full ownership of a kpi last month and spent more time fixing its work than doing it myself
@TheGeorgePu interesting. couldn't find anything on perplexity computer charging per task like that. their api is usage based but way cheaper per query. is this some new agentic feature or you got hit with a bug
@rezoundous yeah apparently there's a prompt caching bug inflating usage by 10-20x. some people rolled back to version 2.1.34 and it helped. might be worth trying while they sort it out
@RomainCerezo@agazdecki this is why i built @brandjetai to surface where people already talk about your stuff. no ads, just listening to conversations happening in real time. distribution through awareness basically
@Prathkum writing code was never the bottleneck honestly. scope creep, unclear requirements, and getting distracted by the next feature instead of finishing the current one. ai just made all of that happen faster
@Codie_Sanchez been building brandjet for 10 months now. some days it's genuinely boring. but the compound effect is real. every week i look back and the thing is noticeably better than it was a month ago