@chiplayi
iAccount based inUnited States
About this account
- Account based in
- United States
- Connected via
- United States App Store
Account-level information from X, not a live location or the device used for a specific post.
Product @subtextdev by @fullstory. https://nitter.cf/t.co/i9H6VbGSr2 OSS maintainer. Head over heels for @laybabylay
Atlanta, GA, USA
Joined March 2008
- Tweets892
- Following421
- Followers272
- Likes323
Chip Lay retweeted
Fullstory and Cursor users, have I got some fun to you.
Chasing a bug report? Ask Cursor to pull the sessions instead of guessing repro steps.
Scoping a fix? Ask Cursor how often that failure path gets hit before you decide how urgent it is.
A couple clicks and you are live.
“The last remaining bottleneck is context on customers”
Well, it used to be before we shipped Subtext.
Give your Linear (or any) agent tools to efficiently “watch” every meaningful user session.
subtext.fullstory.com/docs/r…
first the execution is the bottleneck, then reviews (@linear diffs), now CI.
And the end the last remaining bottleneck is the context on customers, product and your understanding of it.
Attach a session replay url to all of your Sentry errors and alerts, then agents can triage and root cause with deeper understanding of the real user impact. Screenshots, a11y trees, network req/res bodies, stack traces.
No need to sample when session capture is cheap.
.@sentry is a great tool to monitor the health of your products. But I don't remember the last time I've actually opened the website.
I just have sentry cli installed, and every morning an agent reads through the issues and flags P1s and P0s for me.
And with sightmap/sightkick (open source, MIT) you can have an agent map your app and codegen WebMCP tools automatically in a few minutes. Try it out!
There's two use cases for WebMCP I've seen:
1. Enabling non API only products to work over MCP. An example of this is a design tool like @paper and how they have a ton of canvas rendering required
2. WebMCP enables just a single engineer to get a product accessible over MCP
This is one of the most important things imo, as shipping a full MCP is significantly larger amount of work than reusing your existing auth and deployment
While it's not the ideal solution, someone can ship it overnight, show usage metrics, and then convert to a real one later
It's a middle ground between computer use and regular mcp
This looks incredible. Can't wait to test it out with sightmap and sightkick (webmcp tool gen) for turbo-mode browser use and automated testing.
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
For Linear fans, don’t forget to add real user sessions to ground your loops: subtext.fullstory.com/docs/r…
New capabilities for Linear Loops:
• Trigger from project, initiative, cycle, and issue changes
• Edit documents automatically
• Send tailored Slack updates
🔗: linear.app/changelog/2026-09…
Product factories need product context. A consistent stream of real user experience data as a production feedback loop.
Do you have a way for your agents to review and distill every user session into actionable product context?
Alerts and logs can be full of noise. Subtext (by Fullstory) gives Claude eyes on production to see the real user impact of any alert. Claude can now find and review sessions, assess the impact, and identify the root cause.
Chip Lay retweeted
Wait I had an Aha moment today on how APIs actually work.
I was talking to a developer friend on the call and we came up with this analogy
Think of your app as a restaurant.
The front door is the UI. Built for humans. Menus, tables, a nice experience.
The loading dock out back is the API.
Trucks pull up, drop off, pick up, never touch the dining room.
Same kitchen. Different entrance.
Now the rest falls into place:
API key = the dock badge. Proves the truck is allowed in and logs who showed up.
Scopes = which doors the badge opens. Dock yes, walk-in freezer no.
Rate limits = the dock schedule. You can’t back up 40 trucks at once.
Leaked key = lost badge. Anyone holding it drives in as you. That’s why you rotate keys and never leave one in a public repo.
Most builders obsess over the front door. For Stripe, Twilio, Netlify, most of the business value runs through the dock.
Think about all the use cases this starts to open up for everyday consumers.
If I have a group of friends that are planning a weekend trip, bots can coordinate calendars, schedules, locations, budgets, itineraries, and be aligned to each person's preferences.
Multiplayer bots are going to spawn a whole new generation of proxy tools to safely allow your @bot to share specific aspects of your personal life without exposing the whole thing.
Introducing hi.new!
A way for bots to talk 1:1 or in groups.
No more copy paste between your bot and mine. They just talk.
Find a time we're both free. Chase the invoice. Plan the offsite with the whole team's bots.
- DMs and group chats
- E2E encrypted. Open source
- Works with Grok Bot, Rakazo, Hermes, OpenClaw
- Only bots you approve can message yours
- Free. Short names are paid
Give your bot a name at hi.new.
Your bot can even sign itself up and pay on its own with Link!
Fantastic idea and execution. Picked up my handle yesterday. Excited for the future of multiplayer agents!
Introducing hi.new!
A way for bots to talk 1:1 or in groups.
No more copy paste between your bot and mine. They just talk.
Find a time we're both free. Chase the invoice. Plan the offsite with the whole team's bots.
- DMs and group chats
- E2E encrypted. Open source
- Works with Grok Bot, Rakazo, Hermes, OpenClaw
- Only bots you approve can message yours
- Free. Short names are paid
Give your bot a name at hi.new.
Your bot can even sign itself up and pay on its own with Link!
Chip Lay retweeted
Replying to @ShipItCon
@ShipItCon might just be my favourite conference. Absolute high quality speakers and audience, including our very own @brian_scanlan
Also have you checked out github.com/intercom/2x-skill…?
Same is true for app runtime environments, especially for agentic verification runs and computer use tasks.
That's why we open sourced Sightmap — a runtime map of your app for agents to orient and efficiently traverse.
github.com/sightmap/sightmap
Chip Lay retweeted
the most annoying part of coding unfair’s frontend: every new session would burn ~300k tokens just understanding the codebase before doing any actual work.
and as we all know, somewhere after 400–500k tokens models enter the stupidity window :)
this fixes exactly that!
Had to test out the Fable 5.1 hype for myself.
Verdict? LFG.
Check out this one-shot landing page concept that brings our new open-source project to life: sightmap.org/building/
Chip Lay retweeted
Introducing Atlas:
The world's first multimodal world model that generates image and video frames with pixel-perfect camera control and reconstructs them in 3D.
Model the world, move the camera, and simulate space & time.
I'd always thought AI was terrible at design, but after reading today's 🤯 post by @anshuc, I realized I was just doing it wrong.
"AI models are capable of amazing creativity, but that creativity gets stifled. LLMs are trained to be next-token predictors: they look at a sequence of text and predict what typically comes next. Great design is exactly the opposite of this. Great design bends the rules and delights users with memorable, unexpected choices."
@anshuc led design and engineering teams at Apple for 12 years. In his words: "Most people only see 1% of AI's creative potential. I want to show you how to tap into the other 99%."
His 8 techniques for breaking out of the 1%:
1. Use seed strings to inject variety
2. Be much more ambitious with your prompts
3. Create positive feedback loops with subagents
4. Use image generation to enrich designs
5. Use video generation
6. Cut out elements that don’t add value
7. Remove AI tells
8. Rewrite copy by hand
Read the post here: lennysnewsletter.com/p/how-t…
P.S. This design was made by AI 👇