Joe McKenney retweeted
Meet Glyph Cluster on AI Gateway.
๐๐๐๐๐๐๐/๐๐๐ข๐๐-๐๐๐๐๐๐๐
โข Free while in stealth for a limited time
โข Access restricted to paid users
โข Use it for coding & agentic knowledge work
โข Prompts may be used for model improvement
vercel.com/changelog/glyph-cโฆ
Joe McKenney retweeted
I was so impressed with Vercel's AI SDK and Gateway I decided to write something up.
Link is below if you're interested ๐๐
Joe McKenney retweeted
Vercel is the fastest [AI Gateway] infrastructure in the world
Replying to @computesdk
Joe McKenney retweeted
1 line of code in @aisdk saves you 90% or more in DeepSeek v4 Flash AI Gateway tokens
DeepSeek v4 Flash via Novita is 90% off for Vercel Pro till 8/11.
๐๐๐๐๐๐๐๐๐ก๐({
๐๐๐๐๐: '๐๐๐๐๐๐๐๐/๐๐๐๐๐๐๐๐-๐๐บ-๐๐๐๐๐',
๐๐๐๐๐๐๐๐๐พ๐๐๐๐๐๐: {
๐๐๐๐๐ ๐๐ข: { ๐๐๐๐๐: ['๐๐๐๐๐๐'] }
}
})
vercel.com/changelog/deepseeโฆ
Joe McKenney retweeted
Advanced spend budgets now on AI Gateway.
Cap and get alerted on spend by team, project, and API key.
๐๐๐๐๐๐ ๐๐-๐๐๐๐๐ ๐๐ข ๐๐๐๐๐๐๐ ๐๐๐ ๐๐๐๐ \
--๐๐๐๐๐ ๐ป๐ถ๐ถ๐ถ \
--๐๐๐๐๐๐๐-๐๐๐๐๐๐ ๐๐๐๐๐๐๐ข
vercel.com/changelog/ai-gateโฆ
Joe McKenney retweeted
We've been explaining AI Gateway as a Cฬถoฬถnฬถtฬถeฬถnฬถtฬถ Token Delivery Network. Like a CDN for AI models.
One great feature of CDNs is the ability to dynamically re-route or deny traffic without redeployment.
When Fable was suddenly retired, we worried about production workloads depending on it. Even Fable aside, models get retired quite often as GPU capacity is heavily contested. Yet our data shows people still having happy production traffic of older model versions!
We're solving this now with AI Gateway Rules. e.g.: you can now rewrite model "routes" on the fly!
๐๐๐๐๐๐ ๐๐-๐๐๐๐๐ ๐๐ข ๐๐๐๐๐ ๐๐๐ \
--๐๐ข๐๐ ๐๐๐ ๐๐๐๐ \
--๐๐๐๐๐๐ ๐๐๐๐๐๐๐๐๐/๐๐๐๐๐๐-๐๐๐๐๐-๐ป \
--๐๐๐๐๐๐๐๐๐๐๐ ๐๐๐๐๐๐๐๐๐/๐๐๐๐๐๐-๐๐๐๐-๐ป
This adds to the repertoire of capabilities and techniques AI Gateway brings to the table to recover lost tokens. If you drop tokens, you lose revenue and customers ๐ธ
Kudos to @rtaneja_ @crowprose @shaper โ we went from CLI design to production-grade capability very quickly!
AI Gateway now supports routing rules.
Migrate to new models by setting rules in the CLI, no code changes in your app or agent required. vercel.com/changelog/ai-gateโฆ
Joe McKenney retweeted
Heartbreaking, what an articulation @crowprose
crowprose.com/blog/competencโฆ
So after all these hours talking about AI, in these last five minutes I am going to talk about:
Horses.
Engines, steam engines, were invented in 1700.
And what followed was 200 years of steady improvement, with engines getting 20% better a decade.
For the first 120 years of that steady improvement, horses didn't notice at all.
Then, between 1930 and 1950, 90% of the horses in the US disappeared.
Progress in engines was steady. Equivalence to horses was sudden.
Joe McKenney retweeted
5 months ago I broke down Figma's new AI onboarding. They're back with a new iteration! The most interesting part: ๐๐ ๐ฝ๐ฟ๐ผ๐บ๐ฝ๐๐ ๐ฎ๐ฟ๐ฒ ๐๐ต๐ฒ ๐ป๐ฒ๐ ๐๐ฒ๐บ๐ฝ๐น๐ฎ๐๐ฒ๐. An updated breakdown ๐
In the last iteration, an AI input form was presented side-by-side with pre-built usecase templates.
This makes sense, they both solve the cold start problem of helping users go from from 0 โ .5 (they don't get users to 1 yet).
I wondered if, in the future, AI entry points and templates would merge.
With this new iteration, Figma has essentially made templates from pre-written, use case prompts!
Templates are great at presenting use cases (push). Prompting is great at capturing intent (pull).
This experience is starting to have the best of both worlds, where the entry point into prompting is structured and easy to understand, but I can still edit the prompt to fit my intent.
I like this and I think it points to a strong vision for AI-generated getting-started experiences (rather than just throwing me into an input).
Of course, templates still play a big role for Figma and Figma has a huge community with tons of helpful resources.
But now AI has a much larger in-product entry point than templates.
I could imagine other tools taking a similar approach!
Joe McKenney retweeted
Another incredibly polished Dopt powered @Superhuman product announcement went out today
Shoutout to primary CTAs that actually enable the thing instead of taking you on a wild goose chase through settings pages
Joe McKenney retweeted
The new and improved universal growth loop:
1) build a great product
2) get Reforge on board as a customer
3) post about Reforge being a customer
All jokes aside, as long time fans of their content, we couldn't be more excited to be working with the team at Reforge.
Joe McKenney retweeted
Really awesome example from Zapier of using AI to help users troubleshoot errors. This is a great demo of the power of contextual AI (no prompt required!)
Joe McKenney retweeted
I'm just going to say it: SOC2 is a cancer in our industry with strong network effects. The fact that it costs $20k and takes non-trivial amount of time yearly for a documentation that in the end has very little to do with actual security.
Don't get me wrong, many of the controls and proposed practices are useful and should be adopted, but it doesn't need to be done in a audit process that costs 5 figures. SOC2 is old school method that should have been re-thought and not implemented as id but here we are; startups are selling to each others and requiring SOC2 cert in return to show they are compliant for future enterprise customers. It's all a cargo cult and good part of it is self-inflicted harm. We should do better, have a low cost monitors without auditors and focus on pen-tests and other actually valuable methods that don't waste time and focus
Joe McKenney retweeted
My 5 year prediction: we'll realize it was a mistake to write all of our high-level tools in a language built around systems-level memory safety (Rust)
Joe McKenney retweeted
We just launched a new Dopt example: ๐ญ๐ก๐ ๐จ๐ง๐๐จ๐๐ซ๐๐ข๐ง๐ ๐ก๐ฎ๐!
โ
Checklist
โ
Tours
โ
Help & Resources
โก๏ธ Seamlessly integrated into your product and at your user's fingertips.
Check out the best hub examples from @JuneDotSo, @framer, @HubSpot, and @remote โฌ๏ธ
Joe McKenney retweeted
We're doing a โจ ๐๐๐ช๐ฃ๐๐ ๐๐๐๐ โจ next week! As a part of it, we're going live on @ProductHunt next Wednesday, November 8th!
Dopt Analytics will give you out-of-the-box, analytics of the product onboarding, education, and announcements you build with Dopt.
Joe McKenney retweeted
Shout out to @AutoblocksAI for shipping this ๐ฅ onboarding hub and getting started checklist with @doptcom!
I love how it's a centralized place where users can find the next best step and other things like docs โจ
Been a blast working with @haroonc and @nolte_adam ๐
Joe McKenney retweeted
๐ Announcing ๐๐ก๐ ๐ฎ๐ฅ๐ญ๐ข๐ฆ๐๐ญ๐ ๐ฉ๐ซ๐จ๐๐ฎ๐๐ญ-๐ฅ๐๐ ๐จ๐ง๐๐จ๐๐ซ๐๐ข๐ง๐ ๐ฉ๐ฅ๐๐ฒ๐๐จ๐จ๐ค โ a comprehensive guide to building better onboarding made with PLG industry leaders from Notion, Superhuman, Loom, and Miro. โ
Joe McKenney retweeted
Earlier this week we at Dopt had a blast hosting a dinner with founders and executives from companies like Notion, Superhuman, Clockwise, and Stych to chat about all things PLG ๐
A few takeaways โ
Joe McKenney retweeted
โจ Introducing @UnrevealedHQ 1.0 โจ
After almost 1 year in the making, Unrevealed is out of beta! And we are live on Product Hunt ๐ป producthunt.com/posts/unreveโฆ
How it started ๐ฃ
A year ago, I was leading the Developer Experience at Maze, trying to improve the life and productivity of my fellow Engineers.
It kept bugging me how much complexity was involved when answering a simple question:
โCan this user use feature X?โ ๐ค
The complexity didnโt lie so much in the code itself, but in the fact that, most often than not, developers and non-technical stakeholders had to collaborate on this problem.
The tools weโre using to solve it are powerful, but donโt address that fact.
It shouldnโt require a Computer Science degree, pages of documentation, and 10 different steps to add a customer to a beta, run an A/B test, or simply check that a feature is enabled for a customerโฆ but thatโs what it feels like.
For years, we've been using feature flags to solve these problems. And theyโre a great tool! But without the right data models around them, they're still a pretty primitive technology.
I wanted a tool with the power of feature flags, but that anyone could use. I couldnโt find anything out there, so I decided to build it myself. Unrevealed was born ๐ถ
How it's going ๐ช
We started the private beta in April. After a few months of improvements to the product, and a summer spent bringing a whole new version to life, we are proud to call what we have built Unrevealed 1.0 ๐ฅ
Unrevealed is not just a tool for developers, anyone can use it:
๐
โโ๏ธ Youโll never have to โconfigureโ a feature flag in Unrevealed ย ย ย ย
๐ Features simply get from one stage to the next when theyโre ready ย ย ย ย
๐ You won't see any complex rules that don't matter for the sage the feature is in
But I didnโt leave a role in DevEx to stop caring about it, so weโve got devs covered too:
๐ช Our SDKs are so simple they feel like magic ย ย ย ย
๐
Theyโre fully typed thanks to our code generator ย ย ย ย
๐๏ธ We have a Toggler component to switch features in dev or your preview environments ย ย ย ย
๐จโ๐ป Youโll get context about your features without leaving VSCode, thanks to our extension
Help us spread the word ๐
We're Live on Product Hunt Today, if you want to support us, you can start by doing it there ๐ producthunt.com/posts/unreveโฆ
Also leave a comment to let us know what you think, and repost this so more people can see it ๐


























