that is one of the most unhinged landing pages i've ever seen, it looks like a vibe coded claude mess that somehow raised 100M
is no one looking at what they're releasing anymore? do people just not care? i guess they don't have to if some intern can just prompt "make cool website for our useless company" and then 100M appears at your door from eager VCs
I think the hardest thing for me is all these sites and services coming out with phrases like: “Trust and govern every answer” or “Bring BI to where people work” or “self-serve in plain language” or “Trust and control behind every answer”
I’m like, what the f*k. We’re busy doing this shit for a little-hundred-mil?
Do we REALLY need these little guide notebooks? Everyone has AI just, I dunno, stuff the shit in SQLite and generate a graph.
One more thing: how many f*king insights do companies need? This shit is like analysis paralysis.
Keep it simple people, don’t pay for this crap. If you need to gather all your little magical documents, maybe it’s time to consider cleaning out the attic instead.
the best part is how they don't help at all for the hard part (collecting and curating high quality data) and instead help with the part any intern with python and a stem undergraduate could have done already (make a pretty graph)
armed with the free version of chatgpt, pretty much anyone capable of understanding or composing the question can summon such a graph nowadays, given the data.
Half of the boxes overlap at least on Safari, it's actually just broken.
The Show NexaCorp Q3 sales box, Slack box and Hex CLI box is half cut off on the right so you can't click the Codex tab, then scrolling down the "vability" which I assume is observability is cut off to the left, and the "adiness" box overlaps "Flexible by design” which I guess fair enough lol.
Worst of all they committed the cardinal sin of hijacking scrolling. Please for the love of god do not hijack scrolling.
Definitely unhinged, but in a somewhat interesting way. It's like they made "AI slop" into a design language and decided to play with it. You can easily tell when a website is low-effort one-shot Claude output, and this isn't that.
I liked Hex. I introduced it to my company and now it has become the main BI tool.
However, since they started pushing the whole AI thing I no longer like them.
It may speed up doing analytics, but makes everything hard to maintain, creates a ton of duplication and adds yet another layer to take care of.
I also think AI may come back to bite them in the ass. What's the point of an analytics platform is you can create dashboards from scratch directly using AI. What are they adding to the table?
They should have stuck to just making developers lives easier instead of that weird objective that is trying to make analytics accessible for everyone
Its the same conundrum any BI tool is going to have (sigma, tableau, redash)
likely they focus on something orthogonal, like trust and security, permissions, etc
fwiw i work at one of their 'top customers' and found their ai to be really useful for spinning up basically ephemeral dashboards for anyone to use (aka, i dont need to file a request to the data team and wait for someone to get back to me or go back and forth). i just ask it to do the thing.
The reliable layer that captures the data. That is what they bring to the table. For now.
Once AI can orchestrate and maintain a project with dozens of clients without slipping, they will be relegated to non-technical SMBs - many enterprises would just build their own shit. No contracts to sign. Full control. AI hosts and maintains. Great.
Then they might pivot to security (heck I would if I were in any SaaS business today).
> Once AI can orchestrate and maintain a project with dozens of clients without slipping, they will be relegated to non-technical SMBs - many enterprises would just build their own shit. No contracts to sign. Full control. AI hosts and maintains. Great.
The problem with this is that, as of today, no AI company will give you support if the AI generated product doesn't work, at least no company I know of. When we pay for a software product, we mostly pay for the support and accountability transfer to the provider.
I mean at least that's what we used to do. Nowadays support and software quality are so bad that maybe one should just vibe code everything. So I think I've come around to your opinion while writing this comment lol.
We are heavy Hex users at OpenRouter. I was skeptical that it would replace "core" reporting tools like Looker, but the efficacy of the Hex agent has made it our go-to for just about all of our analytics, and allowed us to get incredibly far with a single (very good) person on the data team.
We can make completely reproducible reporting when we need it (for e.g. our weekly business reviews), while the agent can answer all of the ad-hoc stuff that would normally be done by a small army of analysts scurrying around.
Writing has been on the wall since before LLMs took over. Companies invested an incredible amount of money into "big data" over the past 15 years, and the returns are still nowhere to be seen. Data science and data engineering departments have been among the most impacted in all the layoff waves since covid.
It was always a pointless job. I’d watch data people manually put together PowerPoints by hand when it’s the same powerpoint every week and could be entirely automatic.
But they don’t, because they want to look busy. And it’s not like they actually understand the data anyway, because they’ve never worked in the area the data comes from.
> And this was leading to some major issues: our agent harness was becoming bloated with hundreds of tools, which were required to teach agents about how to use Hex-something emphatically out of distribution. Our compatibility with external agents, because those products were having to use Hex through a keyhole via MCP versus being able to work with something more native. And our compute engine wasn’t built to support the high-speed, high-concurrency, iterative workflows agents seem to like.
I'm having déjà vu... I read almost the same thing word for word ~6 months ago. I think it was from an open source project.
Edit: February, but it was published in November:
What I learned building an opinionated and minimal coding agent
Actually I think I've seen the same thing three or four times this year. But yeah, actually solving the actual problems you actually have is a good idea :)
Cool. But honestly all you need is mcp, a semantic layer and let your team run their analysis. Then use something like evidence for bi reporting via duck db. Costs almost nothing.
As a data generalist in an business, it makes zero sense for me to spend my time building a it tool like Hex when I can buy one and spend my time actually moving the needle on business outcomes.
I used to use Hex but Marimo (OSS local first) honestly renders it completely obsolete for me
It similarly offers the combo of SQL and dataframes, built in dataviz and UI widgets, great LLM integration and essentially everything else one would want from either a notebook or a BI tool
How do you share a dashboard built in Marimo with someone that doesn't have any coding experience (for example someone from marketing)
Can you host it somewhere?
This is getting crazy and is another feather in the cap of unhinged American investing and business culture right now.
My first question, what does a BI company need $100M for?
Why does it seem that Figma’s logo is on every company’s trust badge list. Do they spend millions on SaaS subscriptions? Or are they getting bamboozled into getting their name hijacked for trust?
Is this just me or do other people feel this is getting out of hand?
Yes, Claude + skills is what we use too. Big picture, all of the old BI players are living on borrowed time. If you have to drag and drop; you'll be dropped.
In my company, we're pretty happy with Claude plus Metabase CLI (self-hosted). We quickly get any dashboard we want, regardless of the skill level of the person asking.
This site is really a headache... Scrolling and trying to read through the endless content straight up isn't pleasant, which is really hard to achieve.
God, what sort of a world do we live in where curiosity has to be "unlocked". No. Curiosity is the thing that happen when you naturally want to know something. If it has to be unlocked, it's not curiosity. And even if there were some meaning one could ascribe to unlocking curiosity, it wouldn't be with some sort of vibe-coded behemoth AI analytics system....
Glad I got out of the corporate world as soon as I could. It's all rather inhuman.
It means that Hex and their suppliers are themselves not in contravention of the MSA, and that they promise they are diligent in screening for compliant labor practices.
For a SaaS, it's more or less just compliance-flavored magic words.
I used to make hardware products and it was very hard to sell in to bigger UK and Australian businesses until we could certify here.
Downvote me if you want, I'm flaunting the guidelines in service of the discourse.
If I may posthumously discipline the horse - when did the suffixes "-native" and "-first" merge in unholy matrimony?
Was it when B2B marketing realized that engineers equated "native" with "best/fastest/simplest solution available"? Or is it even more stupid than that?
Idk. Honestly native is a good descriptor if you mean “we designed it explicitly for ___” as opposed to “we realized last quarter that ___ was the new cool thing, so we bolted it awkwardly onto our product”
I also hate them, along with “go to market” and “customer focused”.
I like your origin idea, but “cloud-native” is practically an oxymoron. Somehow taking a simple native application and perverting it to the cloud with added layers of abstraction in both the client and the server is somehow “cloud native”?
Makes downtown New York seem like “natural living”.
> https://hex.tech/
that is one of the most unhinged landing pages i've ever seen, it looks like a vibe coded claude mess that somehow raised 100M
is no one looking at what they're releasing anymore? do people just not care? i guess they don't have to if some intern can just prompt "make cool website for our useless company" and then 100M appears at your door from eager VCs
I think the hardest thing for me is all these sites and services coming out with phrases like: “Trust and govern every answer” or “Bring BI to where people work” or “self-serve in plain language” or “Trust and control behind every answer”
I’m like, what the f*k. We’re busy doing this shit for a little-hundred-mil?
Do we REALLY need these little guide notebooks? Everyone has AI just, I dunno, stuff the shit in SQLite and generate a graph.
One more thing: how many f*king insights do companies need? This shit is like analysis paralysis.
Keep it simple people, don’t pay for this crap. If you need to gather all your little magical documents, maybe it’s time to consider cleaning out the attic instead.
the best part is how they don't help at all for the hard part (collecting and curating high quality data) and instead help with the part any intern with python and a stem undergraduate could have done already (make a pretty graph)
armed with the free version of chatgpt, pretty much anyone capable of understanding or composing the question can summon such a graph nowadays, given the data.
I think it looks cool.
The massive lag and jitter as I try to scroll down doesn't feel very cool. It somehow gets better a few screens down, but still...
Also it takes a few screens to figure out what is it even that they do.
scrolls fine for me in an iPhone.
It doesn't on Firefox
Half of the boxes overlap at least on Safari, it's actually just broken.
The Show NexaCorp Q3 sales box, Slack box and Hex CLI box is half cut off on the right so you can't click the Codex tab, then scrolling down the "vability" which I assume is observability is cut off to the left, and the "adiness" box overlaps "Flexible by design” which I guess fair enough lol.
Worst of all they committed the cardinal sin of hijacking scrolling. Please for the love of god do not hijack scrolling.
The way it's slightly off-center is unpleasant
It’s flashy, but nearly completely devoid of any substance.
It's like they asked an AI to throw together every possible flourish into a single page.
I dunno. Maybe the value is in the data siphon from gullible companies trusting Hex with their data.
Do you think that they raised the money based only on this website?
Here's a quote from TFA
> See, we were founded in 2019, making us a pre-AI company (ew!).
I'd wager the website is just the most obvious symptom of the deeper problem.
It's the generic tech landing boiler plate page with a very poor color palette. Outside of the colors it's looks very non offense.
What's the issue?
Darn, i liked their old site better
I’ve seen this terrible mix of sans serif and serif italics in titles from /frontend-design one too many times. What the hell is with that?
the model didn't just invent those design elements without training data.
Definitely unhinged, but in a somewhat interesting way. It's like they made "AI slop" into a design language and decided to play with it. You can easily tell when a website is low-effort one-shot Claude output, and this isn't that.
As much as I hate to say it, did anyone ever care beyond a few employees? Get rid of them and well..
I liked Hex. I introduced it to my company and now it has become the main BI tool.
However, since they started pushing the whole AI thing I no longer like them.
It may speed up doing analytics, but makes everything hard to maintain, creates a ton of duplication and adds yet another layer to take care of.
I also think AI may come back to bite them in the ass. What's the point of an analytics platform is you can create dashboards from scratch directly using AI. What are they adding to the table?
They should have stuck to just making developers lives easier instead of that weird objective that is trying to make analytics accessible for everyone
Its the same conundrum any BI tool is going to have (sigma, tableau, redash)
likely they focus on something orthogonal, like trust and security, permissions, etc
fwiw i work at one of their 'top customers' and found their ai to be really useful for spinning up basically ephemeral dashboards for anyone to use (aka, i dont need to file a request to the data team and wait for someone to get back to me or go back and forth). i just ask it to do the thing.
The reliable layer that captures the data. That is what they bring to the table. For now.
Once AI can orchestrate and maintain a project with dozens of clients without slipping, they will be relegated to non-technical SMBs - many enterprises would just build their own shit. No contracts to sign. Full control. AI hosts and maintains. Great.
Then they might pivot to security (heck I would if I were in any SaaS business today).
> Once AI can orchestrate and maintain a project with dozens of clients without slipping, they will be relegated to non-technical SMBs - many enterprises would just build their own shit. No contracts to sign. Full control. AI hosts and maintains. Great.
The problem with this is that, as of today, no AI company will give you support if the AI generated product doesn't work, at least no company I know of. When we pay for a software product, we mostly pay for the support and accountability transfer to the provider.
I mean at least that's what we used to do. Nowadays support and software quality are so bad that maybe one should just vibe code everything. So I think I've come around to your opinion while writing this comment lol.
We are heavy Hex users at OpenRouter. I was skeptical that it would replace "core" reporting tools like Looker, but the efficacy of the Hex agent has made it our go-to for just about all of our analytics, and allowed us to get incredibly far with a single (very good) person on the data team.
We can make completely reproducible reporting when we need it (for e.g. our weekly business reviews), while the agent can answer all of the ad-hoc stuff that would normally be done by a small army of analysts scurrying around.
OpenRouter only having 1 person in the data team is completely nuts. A canary in the coal mine for whats to come in the whole data space :|
1 or 2 exceptional ICs, a clean semantic layer and agents is "all you need".
RIP to the sexiest job of the 21st century ...
Writing has been on the wall since before LLMs took over. Companies invested an incredible amount of money into "big data" over the past 15 years, and the returns are still nowhere to be seen. Data science and data engineering departments have been among the most impacted in all the layoff waves since covid.
As far as I know OpenRouter has <100 employees, maybe even <50.
1-2% headcount on data is below average for a tech company but not unbelievable either.
It is true you can get a lot further than before with "just" a clean semantic layer though!
The secret here is that building the semantic layer is a ton of work, predominantly done by data people.
Far less promising companies with zero profits often have thousands more employees…
It was always a pointless job. I’d watch data people manually put together PowerPoints by hand when it’s the same powerpoint every week and could be entirely automatic.
But they don’t, because they want to look busy. And it’s not like they actually understand the data anyway, because they’ve never worked in the area the data comes from.
thanks for sharing.
any other tools that are also key to your stack that you wouldn't mind sharing, so others can learn from and try out?
Love to learn more how to become a very good data person, mind sharing more about it.
> And this was leading to some major issues: our agent harness was becoming bloated with hundreds of tools, which were required to teach agents about how to use Hex-something emphatically out of distribution. Our compatibility with external agents, because those products were having to use Hex through a keyhole via MCP versus being able to work with something more native. And our compute engine wasn’t built to support the high-speed, high-concurrency, iterative workflows agents seem to like.
I'm having déjà vu... I read almost the same thing word for word ~6 months ago. I think it was from an open source project.
Edit: February, but it was published in November:
What I learned building an opinionated and minimal coding agent
https://news.ycombinator.com/item?id=46844822
Actually I think I've seen the same thing three or four times this year. But yeah, actually solving the actual problems you actually have is a good idea :)
Cool. But honestly all you need is mcp, a semantic layer and let your team run their analysis. Then use something like evidence for bi reporting via duck db. Costs almost nothing.
Your comment reminds me of this one about Dropbox: https://news.ycombinator.com/item?id=9224
As a data generalist in an business, it makes zero sense for me to spend my time building a it tool like Hex when I can buy one and spend my time actually moving the needle on business outcomes.
Just because you can doesn't mean you should.
I used Hex at $PREVJOB and it was easily one of my favorite tools. It is maybe the one BI tool that I have felt truly attached to in my career
I used to use Hex but Marimo (OSS local first) honestly renders it completely obsolete for me
It similarly offers the combo of SQL and dataframes, built in dataviz and UI widgets, great LLM integration and essentially everything else one would want from either a notebook or a BI tool
How do you share a dashboard built in Marimo with someone that doesn't have any coding experience (for example someone from marketing) Can you host it somewhere?
They have "marimohub" you can share stuff on but also can export to standalone HTML or docs with ``marimo export``
It has WASM support via Pyodide so people can use the notebooks without a python server or toolchain
You can self-host marimohub (BYO compute, object storage, auth, managed AI, etc)
https://github.com/marimo-team/marimohub
[dead]
Funny, both are run by Palantir alums.
This is getting crazy and is another feather in the cap of unhinged American investing and business culture right now.
My first question, what does a BI company need $100M for?
Why does it seem that Figma’s logo is on every company’s trust badge list. Do they spend millions on SaaS subscriptions? Or are they getting bamboozled into getting their name hijacked for trust?
Is this just me or do other people feel this is getting out of hand?
Fascinating. We used to use it quite a bit. But nowadays its just claude + skills.
Yes, Claude + skills is what we use too. Big picture, all of the old BI players are living on borrowed time. If you have to drag and drop; you'll be dropped.
In my company, we're pretty happy with Claude plus Metabase CLI (self-hosted). We quickly get any dashboard we want, regardless of the skill level of the person asking.
I'm surprised at the negative feedback in the comments.
We signed up for Hex a few weeks ago. The PoC was smooth. My colleagues love it.
This site is really a headache... Scrolling and trying to read through the endless content straight up isn't pleasant, which is really hard to achieve.
Is this basically a public version of Meta’s bento notebooks?
Bento is an internal version of Jupyter notebooks.
and can be shared. I know what Bento is. which from what I can is what Hex provides. I left before agents was integrated into Bento (if it was).
> to unlock your team's curiosity
God, what sort of a world do we live in where curiosity has to be "unlocked". No. Curiosity is the thing that happen when you naturally want to know something. If it has to be unlocked, it's not curiosity. And even if there were some meaning one could ascribe to unlocking curiosity, it wouldn't be with some sort of vibe-coded behemoth AI analytics system....
Glad I got out of the corporate world as soon as I could. It's all rather inhuman.
Have....they seen Arcane ?
> "If we didn’t have Hex, we might’ve left more than $100M on the table. That might even be an underestimate."
Hex saved Mercor $100MM. That's impressive!
I've never seen a Modern Slavery Act statement, does that mean no 996?
It means that Hex and their suppliers are themselves not in contravention of the MSA, and that they promise they are diligent in screening for compliant labor practices.
For a SaaS, it's more or less just compliance-flavored magic words.
I used to make hardware products and it was very hard to sell in to bigger UK and Australian businesses until we could certify here.
When I see one of these in a contract, I mark it up to also include the 13th Amendment.
Downvote me if you want, I'm flaunting the guidelines in service of the discourse.
If I may posthumously discipline the horse - when did the suffixes "-native" and "-first" merge in unholy matrimony?
Was it when B2B marketing realized that engineers equated "native" with "best/fastest/simplest solution available"? Or is it even more stupid than that?
Idk. Honestly native is a good descriptor if you mean “we designed it explicitly for ___” as opposed to “we realized last quarter that ___ was the new cool thing, so we bolted it awkwardly onto our product”
Native already had very specific meaning to technologists. Seems to me the industry is trading on the existing credibility of "things that are native"
I also hate them, along with “go to market” and “customer focused”.
I like your origin idea, but “cloud-native” is practically an oxymoron. Somehow taking a simple native application and perverting it to the cloud with added layers of abstraction in both the client and the server is somehow “cloud native”?
Makes downtown New York seem like “natural living”.
We built a much simpler alternative to this over complicated mess that is Hex.
If you're interested in using it, you can 1-click deploy.
https://dbpro.app/studio
No python notebooks in dbpro, that’s a nonstarter for comparison. Different product.