Tyler Postle

@droidbuilds they're cooked

MrBeast plans to trap 1000 vibe coders in a room without Claude first person to center a div manually wins $1 million

@thekevinqi @Voker_ai @kytsune @SiliconANGLE @theCUBE The first bite was… can’t say the same about the next 24

@PostleTyler @Voker_ai @kytsune @SiliconANGLE @theCUBE were the pies tasty at least?

Huge week for @voker_ai - Ranked top 10 on Product Hunt - 70k LinkedIn impressions in the last week - 10k website visits - 60 customers onboarded - Written up by @kytsune at @SiliconANGLE & @theCUBE - Got 25 pies thrown in my face - Threw a RAGER to celebrate our new Agent Analytics Platform with 150 of our supporters, customers and friends - Hosted events at Lightfield and Corgi Cafe - Met up with new customers at ProductTank San Francisco, Cursor and WorkOS Appreciate all of you.

Had the pleasure of hosting an event at Lightfield on Monday, with CTO's @vesko_st, Adit Sanghvi (from @getdecisional), and @bertie_ai, to talk about what it takes to go from shipping your first agent to being the best in your category.

@Moustafa_Awad @ycombinator @FundersClub Context is key! Monitor and optimize your agents or fall behind!

@PostleTyler @ycombinator @FundersClub This is exactly the gap. Agent teams do not just need logs. They need a way to tell whether the agent made the right operating decision in context.

Today, I'm thrilled to disclose Voker’s $2.2M in pre-seed funding from @ycombinator and @FundersClub. I’ve watched product teams ship AI agents with zero visibility into whether they're actually working. Logs don't tell you if your agent is helping users. You can't review 10,000 conversations manually. And metrics don't appear on their own. That's why we built Voker.

@Roger0975587409 @Voker_ai Already launched yesterday, but happy to check it out! Good luck

@PostleTyler @Voker_ai Congrats on the launch! We're launching tomorrow on PH too, would love to support each other. Happy to rally our team to check out your product if you're open to doing the same for us tomorrow 🙌

The proof is in the pie. Yesterday @voker_ai launched on Product Hunt, and we finished in the top ten. As promised... for every 10 upvotes I got pied in the face.

@guilippert_v4 @ycombinator @FundersClub Or whether you dont need to fix at all! Maybe its not a common pattern, just an edge case that isnt even worth an eval

"You can't review 10,000 conversations manually" is the key line There's a missing layer between logs and metrics: reviewable evidence Raw traces are too much dashboards are too late What teams need is a way to turn a few conversations into cases: what broke, why it broke, and whether the fix belongs in prompt, retrieval, policy, or product

Most AI product teams are flying blind. They know how many agent conversations happened, but they don't know if any of them actually worked. @daltonc(CoFounder @standard_cap, Partner Emeritus at @Ycombinator) put it simply: "You can't improve what you can't measure." It's why we started @voker_ai

@luckyPipewrench @svpino seriously! evals do a good job of catching those "clicked a wrong button" or "api failed" but those complexities of '6 REASONABLE steps' that turn out ot be actually unreasonable.... those are the bugs that take days to figure out without analytics patterns

I like this. The failure mode with agents usually isn’t “a user clicked the wrong button.” It’s the agent took six reasonable-looking steps, picked the wrong tool on step four, retried it twice, and now nobody knows whether the issue was the prompt, the tool, the data, or the app.

How to enable full observability and automatic analytics for your LLM-based application. It takes one library + one line of code, and you get a ton of information for free. This is a no-brainer.

@bygregorr @svpino gotta shoot our shot!

@svpino ngl vendor-sponsored "no-brainer" is still a vendor pitch

@alphabatcher @svpino he's the best isn't he? been following him for some time now, he's the only account I have on push-notifications on my phone haha

@svpino Santiago always give us alpha research as usual

@praveenkoka @svpino exactly! Obs tools are great for technical debugging - but if you have real scale and want insights without tons of configuration, you need purpose built analytics!

@svpino One line of code to enable observability, infinite lines of config to make it actually useful.

@PavanKumarNY @ycombinator @FundersClub Thank you!

@PostleTyler @ycombinator @FundersClub Huge news! Love it

@ycombinator @FundersClub Huge thanks to our investors, employees, partners, and customers. If you're building with AI agents and want more signal, less noise, try Voker for free up to 2000 events/mo: voker.ai and check us out on Product Hunt today! producthunt.com/products/voker…

Voker is the agent analytics platform for AI product teams. It’s simple: just install our llm stack-agnostic SDK and Voker tracks your agent performance and usage data, alerts you when things drift, and helps every stakeholder understand how your agent is really performing. We're at the beginning of understanding how to build AI products that actually work at scale. @voker_ai is how teams get there.

Have you ever wanted to pie a CEO in the face? Tomorrow’s your chance! Watch below to learn how. 🥧

@iliakakhadze @alexrudolphdev The current substitute is vibe-y. Ppl take traces from obs tools and ask ChatGPT to analyze trends (LLMs cant do math or stats). Its a plain old data eng problem - which is what we do

@iliakakhadze @alexrudolphdev Analytics SDK to monitor your AI agents in production.

At a YC hackathon rn. The YC-backed company sponsoring it their site goes down. Been down for over an hour. Some participants walk up to help. They ask what they uses for caching and the founder responds, "I don't know." This is what happens when you vibecode an entire product. How TF are these companies getting into YC is my question. Like c'mon bro, at least know your own product.