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?
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.
I liked Hex. I introduced it to my company and now it's 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 it 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 live easier instead of trying to make analytics accessible for everyone
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.
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.
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”.
> 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 :)
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
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 a low-effort one-shot Claude output, and this isn't that.
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.
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.
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.
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.
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.
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 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?
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.
55 comments
[ 0.20 ms ] story [ 7.5 ms ] threadIt 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
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 it 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 live easier instead of trying to make analytics accessible for everyone
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).
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.
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.
If you're interested in using it, you can 1-click deploy.
https://dbpro.app/studio
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?
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”.
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 :)
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
Also it takes a few screens to figure out what is it even that they do.
> 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.
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.
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.
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.
any other tools that are also key to your stack that you wouldn't mind sharing, so others can learn from and try out?
1 or 2 exceptional ICs, a clean semantic layer and agents is "all you need".
RIP to the sexiest job of the 21st century ...
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!
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.
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.
Hex saved Mercor $100MM. That's impressive!
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?
We signed up for Hex a few weeks ago. The PoC was smooth. My colleagues love it.
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.