Barry McCardel
3.8K posts
Barry McCardel
@barrald
co-founder and CEO of Hex (https://t.co/hbgguInF1h / @_hex_tech) - former @PalantirTech @formationbio - personal site: https://t.co/c38nDG5Dfl
San Francisco, CA Katılım Ağustos 2011
988 Takip Edilen3.8K Takipçiler
Not enough people are talking about how much AI is impacting the role of data science.
I was chatting with a DS friend, and he said that most of his team's work now is reviewing half-assed AI data analysis from PMs and engineers. And that 50% of the time, that analysis is wrong.
The role is becoming less fun.
English
when I first heard of @baseten they were basically a competitor
but then I met @tuhinone and I liked him, and when I heard they were pivoting to inference, I was relieved because I didn't want to compete against him
and then he hired @DannieHerz and I was angry because I wish I had thought of it
and now they're a critical partner for us as we embrace our own many model future at @_hex_tech
I'm so happy for all their success and very excited to share what we've been working on with them!
Tuhin Srivastava@tuhinone
English
@n2parko very grateful for all your feedback + partnership Nate!
English
this has rewired the way we work with data
Olivia Koshy@oliviakoshy
introducing in @_hex_tech: Generative Data Apps backed by code, so you can build anything you want! a dashboard, an editorial, a customer prezo - you name it, you can build it 💜 integrated with your data team's context + built w/ governance, security and observability top of mind
English
The most wondrous and terrifying part of 2026 is the opportunity to re-invent whole swaths of @_hex_tech.
We're in the process of throwing away features we spent years on – and deeply replacing agents.
Generative data apps are a first step – with much more coming soon!
Olivia Koshy@oliviakoshy
introducing in @_hex_tech: Generative Data Apps backed by code, so you can build anything you want! a dashboard, an editorial, a customer prezo - you name it, you can build it 💜 integrated with your data team's context + built w/ governance, security and observability top of mind
English
@anshublog @prukalpa @Snowflake @databricks @Workato Oh Foundry Catalog is more equivalent of like a database catalog i.e., manages table metadata
In any case “unbundling Palantir” is an interesting angle on what you’re doing – I’d argue the value of Palantir is the bundle!
English
@barrald We are the data control layer - security and privacy - as data flows across datastores, models, and agents.
The catalog is replaced by my friend @prukalpa.
The data layer is @Snowflake or @databricks.
The orchestration is @Workato etc.
English
Unbundling Palantir.
Bundling and unbundling. This is the way.
English
@anshublog yeah their diagram is really vague too!
So Skyflow replaces Foundry Catalog? Or Gatekeeper the underlying ACL system? Or something else
English
@barrald It’s their diagram, first and foremost. So maybe start with that.
And then I can explain the substitutions and evolution.
Here’s a link to the source. :)
palantir.com/docs/foundry/p…
English
In 2026, all:
Loops are “agents”
Backends are “engines”
UIs are “canvases”
Apps are “context layers”
Engineers are “members of the technical staff”
Collections of markdown files are "moats"
What am I missing?
English
English
@hwchase17 @isidoremiller I told @barrald "I can't wait to show up to work tomorrow and pass off Izzy's thoughts as my own"
English
🎙️Introducing Max Agency
Max Agency is a new podcast where we go deep on how the best agents are actually being built: architecture decisions, tradeoffs, evals, and everything in between. Each episode, I sit down with engineering leaders who are doing this work in production.
Our first episode features Izzy Miller (@isidoremiller), AI Engineer at Hex (@_hex_tech). Hex has been shipping data agents since before most teams were even thinking about them, starting with single-cell text-to-SQL and graduating to a full Notebook agent that can work autonomously for 20 minutes on a complex analysis.
Izzy has a lot of perspective on what it actually takes to get agents working well in production, and what breaks along the way.
A few takeaways from our conversation:
- Keep your eval sets small enough to hold in your head: Izzy runs 30-50 handcrafted "traps" with multiple repetitions, rather than hundreds of variants. If you can't explain why your agent fails each one, your eval set is too big
- Day zero performance is almost irrelevant: The more interesting question is how the agent compounds. Izzy is building a 90-day simulation where the warehouse evolves and the agent has to accumulate understanding
- You can catch agent errors without seeing the raw outputs: By running an LLM-as-a-judge over production usage and clustering the results, you can surface places where something likely went wrong, without needing to read individual conversations
Watch the full episode on:
- Youtube: youtube.com/watch?v=Xyh1Eq…
- Apple Podcasts: podcasts.apple.com/us/podcast/how…
- Spotify: open.spotify.com/episode/1BJlg3…
YouTube
English
Izzy is a singularly talented AI thinker and tinkerer, and this interview makes me heart-swellingly proud of all the work he and the team are doing at @_hex_tech 💜
Harrison Chase@hwchase17
🎙️Introducing Max Agency Max Agency is a new podcast where we go deep on how the best agents are actually being built: architecture decisions, tradeoffs, evals, and everything in between. Each episode, I sit down with engineering leaders who are doing this work in production. Our first episode features Izzy Miller (@isidoremiller), AI Engineer at Hex (@_hex_tech). Hex has been shipping data agents since before most teams were even thinking about them, starting with single-cell text-to-SQL and graduating to a full Notebook agent that can work autonomously for 20 minutes on a complex analysis. Izzy has a lot of perspective on what it actually takes to get agents working well in production, and what breaks along the way. A few takeaways from our conversation: - Keep your eval sets small enough to hold in your head: Izzy runs 30-50 handcrafted "traps" with multiple repetitions, rather than hundreds of variants. If you can't explain why your agent fails each one, your eval set is too big - Day zero performance is almost irrelevant: The more interesting question is how the agent compounds. Izzy is building a 90-day simulation where the warehouse evolves and the agent has to accumulate understanding - You can catch agent errors without seeing the raw outputs: By running an LLM-as-a-judge over production usage and clustering the results, you can surface places where something likely went wrong, without needing to read individual conversations Watch the full episode on: - Youtube: youtube.com/watch?v=Xyh1Eq… - Apple Podcasts: podcasts.apple.com/us/podcast/how… - Spotify: open.spotify.com/episode/1BJlg3…
English
this is just our first iteration of the Hex CLI. much more coming, so let us know what else you want next!
Barry McCardel@barrald
Agents are hanging out in the terminal, and now @_hex_tech is joining the party 👯🎈 Hex CLI is extremely useful for everything from propagating dbt model updates, to BI migrations, to everything in between It's a new level of malleability, observability, and control for teams – and I'm very excited to keep pushing here! Try it out now: brew install hex-inc/hex-cli/hex
English
And of course the team cooked on the page – you can even use the CLI!
hex.tech/product/cli/
English
Agents are hanging out in the terminal, and now @_hex_tech is joining the party 👯🎈
Hex CLI is extremely useful for everything from propagating dbt model updates, to BI migrations, to everything in between
It's a new level of malleability, observability, and control for teams – and I'm very excited to keep pushing here!
Try it out now: brew install hex-inc/hex-cli/hex
English
Slack’s new AI stuff is somehow both too prominent, and not discoverable enough
Don’t know how they do it
English
@barrald @_hex_tech Feels like an unintentional dig at Cursor lmao
English
Today @_hex_tech is introducing the boldest redesign in the history of software.
As agents transform software, everyone is questioning the role of UX. Will it all become vibe-coded slop? Or will it simply disappear?
At Hex, design is the heart of everything we do – and we believe in beautiful UI as much as ever.
Today we're introducing a preview of Pure Glass – a new language that evolves our UI, replacing harsh borders with frosted edges, and harsh contrast with clear translucency.
No one asked for this – but we're doing it anyway, and are so excited to unleash this on our users this coming fall.
English
If we made /slow mode in Codex, would you use it? What for?
(Slower inference at a cheaper cost)
English
@caitlinbolnick1 @_hex_tech you know I exist only to make you chortle
English
@barrald @_hex_tech Your turtleneck doesn't have enough neck to it Barry
English
during every major tech shift, the entire market commits the cardinal sin of platforms: building the platform before the app
the app informs the platform. always. this is why sales cloud comes before salesforce platform, and why FDEs are app builders, not sales engineers
English