Pinned Tweet
Introducing: The @Every agent, the fastest way to AI-pill your company, now officially out of beta.
With personal agent use, no one can see how their teammates are working with AI. With one “company agent” in @SlackHQ, they can.
The Every agent comes loaded with the skills and workflows we use to run Every, and each person on your team gets $15 in free credits to try it.
Install @Every in @SlackHQ today: every.to/agent?utm_source=x&…
BREAKING:
Introducing the @Every Agent—an agentic coworker in Slack built to help you AI-pill your company.
It comes with:
- Everything you'd expect from a premium company agent. You can delegate difficult, long-running—like opening PRs, writing reports from large datasets, and monitoring social or customer feedback—and get high-quality results all from Slack
- Every's best workflows, pre-installed and always updated. Compound Engineering to Compound Writing. We add more all the time.
- Personalized Frontier Alerts. to help you keep up with new models. When we drop a Vibe Check, the Every Agent will read it and let you know if the new model is relevant to your workflow.
- Zero markup on tokens. What we pay is what you pay—as token prices go down, so does your bill.
You can try it today with $15 in free credits per person, up to $500 per workspace.
After that: it comes with your $30/month Every subscription, plus tokens at cost.
Add the Every Agent to Slack: every.to/agent
Why it's hard to AI-pill your coworkers
If you’re the AI-pilled one at work, you know the problem: you discover something incredible in Claude or Codex, but getting your coworkers to try it is another matter.
The problem is that most of your AI use is invisible. You do it alone, in a terminal or a chat window.
Your coworkers see the finished report or the five new ads, but they never see the request that produced them, or the corrections you made along the way. When you describe it afterward, it sounds like magic, but nobody learns a magic trick without seeing how it’s done.
Why a company agent, in Slack, solves this problem
When you use a company agent like the @every agent you go from prompting in private, to prompting in public.
In a Slack channel, a coworker who would never open a terminal scrolls past a thread with your agent interaction and thinks: I could do that too.
This lets you show instead of tell your coworkers about your discoveries. New AI tools and workflows then spread naturally and virally across your organization.
That's why we made the @every agent for ourselves—and why we're releasing it to you today.
Add the Every Agent to Slack: every.to/agent
Opus 5.5 is our favorite model.
Here's what else the Every team is running, and who splits the big jobs from everything else.
The Daily Driver, in Context Window: every.to/context-window/micr…
12
830
Interesting find from the GPT-6.2 Astra & Sol retest: It keeps using the same grain-generating code.
@tylernishida discovered it when trying to play around with grain levels. But asking for no grain stopped it.
The repeating line is a random-number recipe from a 1992 textbook. Both models even used the same variable names and almost identical cream colours.
Our live re-test of GPT-6.2 Astra & Sol: www.teapawt.org/i/broadcasts/2dKOWNNpN…
1
4
946
BREAKING:
@Google just launched Gemini 4 Neon
We're vibe-checking live right now:
Retesting GPT-6.2 Astra & Sol www.teapawt.org/i/broadcasts/2dKOWNNpN…
6
1,319
Open invitation for anyone at @OpenAI to come on:
Retesting GPT-6.2 Astra & Sol www.teapawt.org/i/broadcasts/2dKOWNNpN…
This week at Every:
• @hammer_mt built map-driven autoresearch, where an agent reads a map of 104 methods and decides which experiments to run (the explainer video is made entirely in code).
• @danshipper built Galley, a word processor that checks each line against his editorial standards as he types.
• @JannikJung had Fable 5.2 design a four-channel amplifier, from choosing the parts to designing the schematic and circuit board.
• @kplikethebird shipped Every Reader, built the Wordbetter word processor, and merged her first pull request.
• The Frontier team published their vibe check for GPT-6.2 Astra and Sol, and @beckyisj and @poojary_yash spun up a live page for the new models.
• @kieranklaassen ported Everycut to Windows and Linux in one day and gave Frontier an RSS feed of all 43 published experiments.
• Andrey Galko updated Every's subscription-gifting page.
• Arielle Shipper walked 15 women through the AI coach she built to reach level eight of Every's AI adoption framework, and @elifnisapolat turned Katie's AI hot takes into a card game.
Read the experiments: frontier.every.to/?utm_sourc…
Retesting GPT-6.2 Astra & Sol www.teapawt.org/i/broadcasts/2dKOWNNpN…
Gemini 4 Carbon found a useful pattern in our survey responses without being asked to look for it.
It grouped readers by how they understood Every: through practical AI use, editorial work, or our products and software studio. Those groups closely matched ones our team had identified independently.
The presentation was harder to read—dense cards inside more cards.
.@hammer_mt describes the trade-off as working with a scientist rather than a management consultant: interesting analysis that can still need help communicating.
Read the full @Google Carbon 4 vibe check: every.to/vibe-check/vibe-che…
1
8
1,216
One of the biggest AI moments in history is Dario Amodei’s 2017 memo, “Big Blob of Compute Hypothesis”.
In it, he outlined a different path for AI: assemble large amounts of computing power, point it at a general problem like predicting the next word, and see whether intelligence emerges.
This was at a time where research was more focused on narrow problems, @kevinroose says.
Watch the full interview with @NataliaZarina:
every.to/podcast/the-inside-…
“You might not want to vibe code a bridge or a building.”
@NotionHQ cofounder @ivanhzhao loves vibe coding—it’s the spirit he started Notion with. But there are still some tools he’d rather use than build himself.
At Thesis: 2027, he joined @danshipper to talk about what to build yourself (or not), and why you shouldn’t “surrender” your understanding of how things work, even as AI is doing the building.
Watch the full conversation:
00:00 Introduction
01:18 Why vibe coding revives the 1960s vision of software
04:40 When people should build their own tools—and when they shouldn't
07:47 Why data and context outlast UI and procedural logic
09:46 The history game Ivan is vibe coding with his wife
12:23 Bringing a 13-year-old company from pre-AI to post-AI
16:00 The TikTokification of AI and its second-order effects
19:46 Don't surrender your understanding to agents
22:01 Why builders need Lego-like blocks to vibe code
28:04 Why the next billion knowledge workers will be agents
Testing every new AI tool is part of @jackcheng’s job. His advice for people with less time: Pick one and learn it deeply.
Get to know how the app works, where it helps, and where it falls short. That familiarity takes time you could otherwise spend switching between tools.
Even our team, which is tasked with testing models, struggles to try everything.
Read the full @Google Carbon 4 vibe check: every.to/vibe-check/vibe-che…
1
812
The Every frontier team is testing Astra 6.2 and Sol 6.2 again today at 3PM ET, live.
Will the models magically be better today than they were during our early preview window? Tune in and find out.
the whole @every frontier team will be re-testing Astra 6.2 and Sol 6.2 LIVE today at 3 PM ET / 1 PM PT!!
@thsottiaux we'd love for you to come on the stream to help us see what we're missing
our mission is to make sure everyone knows where each new model shines (and where it doesn't) so they can use it to do great work. this way, everyone will be able to see how our process works, and we'll be able to show any changes or updates we find!
Every đź“§ retweeted
the whole @every frontier team will be re-testing Astra 6.2 and Sol 6.2 LIVE today at 3 PM ET / 1 PM PT!!
@thsottiaux we'd love for you to come on the stream to help us see what we're missing
our mission is to make sure everyone knows where each new model shines (and where it doesn't) so they can use it to do great work. this way, everyone will be able to see how our process works, and we'll be able to show any changes or updates we find!
Before committing to a new idea, @fredgraver asks @claudeai 3 questions:
• Does this sound like something I would make?
• Is there a real audience for it somewhere?
• What would I need to finish it that I don’t have now?
That last question matters because an exciting premise can mean a long stretch of writing and finding people to help make it. He uses the conversation to decide whether he wants to do that work.
Watch the full episode:
every.to/context-window/ever…
2
775
What’s the most important step when building AI agents for finance teams?
Every’s @baynkr is sharing his experience today at noon ET at Traceable, @chargebee’s virtual AI summit for finance teams.
His session, “What I’ve Learned From Building Agents at Frontier Finance Orgs,” draws on his work as our senior director of AI consulting.
Join us for Ryan’s talk: chargebee.com/traceable/?utm…
Finance teams are busy. They’re building billing agents, automating collections, and figuring out which decisions still need a person.
Today at Traceable, @chargebee is bringing together the people doing that work to share what they’ve built—and what they’re learning.
Every’s @baynkr joins the lineup with lessons from building agents at frontier finance organizations.
The virtual summit starts at 11 a.m. ET today.
Save your spot: chargebee.com/traceable/?utm…
Every đź“§ retweeted
what should we add to the Every Builder Pack for 2027? What are you using to build/operate/grow that everyone needs to be using right now.
every.to/builder-pack
Fable described a raspberry jelly that “never quite stops trembling.”
Gemini 4 Carbon started with “a tall fluted dome of translucent ruby gelatin” and kept going.
In @tylernishida’s comparison, that difference can be spotted through the writing: Fable was more direct and brief; Carbon was more descriptive, technical, and heavy on adjectives.
Read the full @Google Carbon 4 vibe check: every.to/vibe-check/vibe-che…
1
2
1,268
Every đź“§ retweeted
I’m leading a sprint to reboot @every’s Checks product, and I put together a Christopher Alexander inspired project pattern language based on my interviews with the team. Then I made this explainer video to summarize the findings. What a time to be alive
No one on our team loved Astra 6.2.
We have a 30-person team at Every, including our Frontier team, which gets access to these models before they come out, and we put them through their paces on our real work tasks, from coding to writing to design to knowledge work.
This is not a great model. No one on our team who tried it switched back to it.
Astra 6.2 is a real step back for writing. It had strong design defaults but couldn't vary them or be creative, and for coding we found it stopped too soon on end-to-end builds.
Get the full vibe check: every.to/live/openai-astra-s…












