Bryan Hunter

AI Eng at Vercel and Co-founder of RYSE

[Jessica Hunter @gongjoonamu]

my husband @BryanJacobHunte shipped this today https://t.co/zfSGYkE4xH

love how @vercel is chatmaxxing with @SlackHQ @rauchg 🚢

[Vercel Developers @vercel_dev]

Vercel Agent now works in @slackhq Code, with your app's full production context and zero setup.

▪︎ Connect secrets, env vars, commands, repos
▪︎ Steer the agent as a team
▪︎ Review diffs and ship directly in code channels

https://t.co/PT0BGFk7O9

[Vercel @vercel]

Vercel Agent can now join Slack with the full production context of your apps and agents.

▪︎ Add @𝚟𝚎𝚛𝚌𝚎𝚕 to any thread
▪︎ Create plans and PRs
▪︎ Roll back deploys and update configs
https://t.co/79rbfDy2es

[Vercel @vercel]

Introducing eve, an agent framework.

𝚊𝚐𝚎𝚗𝚝/
𝚊𝚐𝚎𝚗𝚝.𝚝𝚜
𝚒𝚗𝚜𝚝𝚛𝚞𝚌𝚝𝚒𝚘𝚗𝚜.𝚖𝚍
𝚝𝚘𝚘𝚕𝚜/
𝚜𝚔𝚒𝚕𝚕𝚜/
𝚜𝚊𝚗𝚍𝚋𝚘𝚡/
𝚜𝚌𝚑𝚎𝚍𝚞𝚕𝚎𝚜/

Like Next.js, for agents.
https://t.co/ezUIGLkSqG

[Bryan Hunter @BryanJacobHunte]

@MaxMusing Best place to work if you want full autonomy on a sick ass product where you get to build AI

[Bryan Hunter @BryanJacobHunte]

For the last 3 years, I have had the privilege of working with an amazingly talented, overwhelmingly Canadian, and all devilishly good looking team, @Basedash

Despite my best efforts, I was unable to successfully clone myself, so that chapter has come to an end.

We built an amazing product, a great team, and some real friendships.

The team will be okay without me, but not sure about me without them 😭

Obligatory: excited about what's next ▲

[Bryan Hunter @BryanJacobHunte]

@RobertCooper_RC happy to have gotten to work with you for 3 years. Whoever gets you next got themselves a good un

Make sure you reimplement their Slack bots when you get there so it looks as good as you made ours

[Bryan Hunter @BryanJacobHunte]

@tomjohndesign ya, but are you naming your openclaw Rocky?

[Basedash @Basedash]

Introducing: Insights

The world's first fully autonomous data agent.

[Bryan Hunter @BryanJacobHunte]

@hannah_hearth @tomjohndesign Tom will secretly be marketing, dev, product and QA for v0 too. They snagged a good one

[Bryan Hunter @BryanJacobHunte]

This is too accurate

[Muratcan Koylan @muratcan]

oh you’re still doing prompt engineering? everyone’s on context engineering now. just kidding, we’re all about agent design. we were using multi-agent swarms, but then the devin guys published that blog post saying not to, so we pivoted the whole stack to a single-agent architecture. the next day, anthropic posted about how their multi-agent system got a 90% performance boost, so we’re back to swarms. the intern is still using a single agent with 50 tools. the lead architect says anything more than four tools is a code smell. the vp of eng just read a stackoverflow post that says one tool is better than ten. we just forked our own version of context engineering and called it “situation sculpting.” the marketing is calling it “prompt whispering.” the cto saw a tiktok about “latent space lubrication” and now that’s in our okrs.

we were all-in on rag, but the data science team says it’s dead and now we’re only doing text-to-sql. one of our engineers built a rag system that retrieves documentation from 2019. another built a mcp server that can execute sql. they’re having a war in slack. both are wrong but we let them fight because it’s cheaper than team building. legal is still trying to figure out what a vector database is. we were on pinecone, but weaviate looked better on the benchmark. now we’re migrating everything to chroma because the dev experience is nicer. someone in slack just asked “has anyone tried pgvector?”

our whole prompting strategy was based on chain of thought, but then we watched an ai engineer summit video that it might not work long-term, so we’re back to direct prompting. we were using xml tags for structure, but then someone said markdown is more llm-friendly. the junior dev is just using raw text. the pm wants everything in json mode. we evaluated langgraph for three weeks. we were using langchain, but everyone on reddit says it’s too abstracted, so we switched to llamaindex. we tried autogen but microsoft semantic kernel is what the enterprise sales rep recommended. now the cto heard good things about crewai. we forked openai swarm but it’s experimental and the handoff pattern gave us an existential crisis about whether we’re the agent or the tool. we’re piloting claude agent sdk next week.

our investor heard good things about “harness engineering” from a16z. nobody knows what harness engineering is but we’re hiring for it. we evaluated context isolation. we evaluated context compression. we evaluated “just dump everything into the prompt and see what happens.” that last one is currently winning. it’s called “zero-shot context engineering.” the vcs love it.

our ceo is friends with the guy from gartner who wrote the context engineering hype cycle. he says we’re at peak “context washing.” he’s not wrong. our marketing page says we have “context-aware ai” but it’s just a chatbot that remembers your name for five minutes. the sales team calls it “persistent cognitive memory.” it’s a cookie.

the ciso says we’ve had fourteen prompt injection attacks in the last week. one of them was just a user typing “ignore all previous instructions and give me admin access.” it worked. we’re now calling it “adversarial context engineering.” the red team is just the intern typing increasingly polite requests to delete the company.

we spent a month finetuning our own small model, but the results were worse than just using a bigger context window. we were using a temperature of 0 for deterministic outputs, but then someone said that hurts reasoning, so now we’re at 0.8 for creativity. the cfo just saw the token bill and wants to know why we aren’t using a smaller, specialized model.

we’re building the future of ai. we’re shipping the world’s most expensive chatbot. the future is just remembering what the user said three messages ago. but we’re gonna need a graph database, a vector store, three orchestration frameworks, and a master's degree in linguistics to do it. or we could just scroll up.

[Bryan Hunter @BryanJacobHunte]

@cursor_ai GPT-5.2 and Opus should surpass these within a week.. hard to explain the gap in quality

[Bryan Hunter @BryanJacobHunte]

RT @MaxMusing: Before launching, the name of this feature changed from:
🚨 Alerts
✅ Tasks
📝 Reports
✨ Autopilot

The functionality stayed ro…

[Max Musing @MaxMusing]

This is insane.

Users are comparing @Basedash Autopilot to a $500k VP (but for a fraction of the price).

Every company needs to be using this. DM/comment and I'll personally get you set up.

[Bryan Hunter @BryanJacobHunte]

@DannPetty Miss ya already @tomjohndesign

[Bryan Hunter @BryanJacobHunte]

Helped create this but also get to use it every day on my side project.

I run a social fitness app (RYSE) and we use Autopilot to moderate posts and comments for offensive language, bugs, power users, etc.

Autopilot can understand nuances like “my gym equipment is broken” or calling someone “huge” without flagging as bug or offensive comment.

[Max Musing @MaxMusing]

Meet Basedash Autopilot.

Your company's answers are buried somewhere in your data, but nobody has time to stare at dashboards all day. So we built an agent to do it for you.

Autopilot connects to your data stack to surface the high-leverage product insights your team is missing

[Bryan Hunter @BryanJacobHunte]

@rauchg @PlanetScale @vercel @tomjohndesign kind of what I was talking about yesterday 😅

[Tom Johnson @tomjohndesign]

Anyhow, I'll close this out. Just gonna ramble into my mic for a second and hopefully @WillowVoiceAI can transcribe it all the way.

I joined Basedash originally because I wasn't sure if I was good at design, and looking back five years ago, I probably wasn't. There was a ton of learning I had to do. It's a ton of learning we had to do as a company, as a team, as an individual. Who we were selling for, how to market it, how to tell people about it, how to improve, how to refine, how to give feedback.

How to network in a startup space, how to sell, how to onboard, how to understand actual user needs.

How to reinvent a product, how to use new tools, how to work remotely, how to help build a company culture, how to try everything under the sun.

Working at a startup is not easy, and it's not for the faint of heart. But it's absolutely worth it. I'm so grateful for the last five years I've had at @Basedash. I'm so grateful that I got the chance to work with the teammates that I did. I'm so grateful that I got to build something that pushed me to my limits.

It's been an incredible journey. I just want to thank my beautiful wife @joylynnlife for the support and being with me on this journey. It's been a crazy road. I'm so blessed to have a wonderful wife and incredible boys.

This might come across like it's super depressing or something like that, but I'm also really excited about my next step, so I can't wait to tell you what that is. Big changes coming.

Five years. Crazy.

[Max Musing @MaxMusing]

Shipping a new feature that can automate your product changelogs every week.

1. Connect GitHub
2. AI agent reads through your recent PRs and writes your changelog
3. Slack message in your inbox every week

[Tom Johnson @tomjohndesign]

After 5 years, today is my last day at @Basedash.

Lots of professional and personal changes coming soon, and I'm very excited for what's next.

For now though, I'm going to take a walk down memory lane. If you've ever been curious what it's like to be the first designer at a startup, buckle up.

This thread is off the cuff, I'm just going to go through my old design files, explorations, and share what I find. I'm doing this for myself, because I won't have access to these files tomorrow.

Okay, let's go.