XCSme

17 ms·

481.
by XCSme 8mo ago
But 80% sounds far from good enough, that's 20% error rate, unusable in autonomous tasks. Why stop at 80%? If we aim for AGI, it should 100% any benchmark we give.
482.
by XCSme 8mo ago
If they don't set the proper UserAgent, it's not trivial. You could add some JS to test it and store it as a tag, maybe using something like: https://stackoverflow.com/a/78629469/407650
483.
by XCSme 8mo ago
Oh, and as for the agents themselves browsing the website, if the agent uses an actual browser with JS enabled (like headless puppeteer), then you would actually be able to see how the agent browsed the website.
484.
by XCSme 8mo ago
I currently do not automatically detect the agents themselves, and most of the bot traffic is ignored, but for the traffic that is not ignored it usually shows as a 0s session (so you can filter for all sessions that have a 0s replay, which
485.
by XCSme 8mo ago
Google banning you for benchmarking is crazy, are you sure that's the cause? How would they even know you are benchmarking?
486.
by XCSme 8mo ago
MiniMax 2.1 didn't really work for my data-parsing tasks, a lot of errors. Instead, this one works surprisingly well for the cost: https://openrouter.ai/xiaomi/mimo-v2-flash
487.
by XCSme 8mo ago
But why only a +0.5% increase for MMMU-Pro?
488.
by XCSme 8mo ago
Their ability to burn through tokens non-stop for hours, days or weeks without intervention.
489.
by XCSme 8mo ago
I feel this. Analytics often ends up being a bunch of charts that don’t really tell you what’s happening. I built UXWizz so I could actually watch how people use my sites with session recordings and heatmaps instead of just staring at table
490.
by XCSme 8mo ago
This trend is wild. I’ve seen so many hype projects where the code barely works and app is buggy as it can be...
491.
by XCSme 8mo ago
430% growth is wild. I built a self-hosted setup to watch session replays and see exactly how people use live demos on my pages [0]. Sometimes users get stuck on a tiny part of the demo, and just fixing that can make a huge difference. Seei
492.
by XCSme 8mo ago
This kind of tracking is insane. I built a self-hosted analytics setup so I can see how my sites are performing without sending anything to ad-tech companies [0]. Keeps the data on my server, gives real insights, and doesn’t contribute to t
493.
by XCSme 8mo ago
Wow, $1,200 for an LED lamp
494.
by XCSme 8mo ago
Offline-first stuff looks awesome. I also built a self-hosted tool for session recordings and metrics to see how clients interact with my sites (UXWizz) Keeps all the data private on my server and gives real insights without relying on thir
495.
by XCSme 8mo ago
Yeah, SaaS tools start feeling bloated fast. I switched to self-hosting analytics so I could track user behavior properly without paying per seat. Was messing around with Grafana and other stuff, but for session replays and metrics I built
496.
by XCSme 8mo ago
AI traffic is really tricky because traditional analytics and SEO tools just don’t capture it. Even basic UTMs often miss the "last hop" behavior from these systems, so it can feel like you have no idea how users actually reach yo
497.
by XCSme 8mo ago
Wait, so this does the same thing the agent would do, but via cloud? What's the advantage?
498.
by XCSme 8mo ago
Maintaining integrations in production can definitely become a hidden burden. Even simple analytics or attribution pipelines can silently fail if APIs or schemas change and that often only shows up when metrics look off. One approach that h
499.
by XCSme 8mo ago
Nice stack. Even just having a way to see core product metrics and user behavior in one place can speed up iteration a lot. Do you self-host PostHog? I built and use instead my own [0]. [0] https://www.uxwizz.com
500.
by XCSme 8mo ago
Really cool landing page, it's so fun to see (not sure if it's AI-generated or not, but looks quite unique).
501.
by XCSme 8mo ago
Love the quick KPI focus. For a bit more insight into user behavior without sending data to external services, Ive been using a self-hosted solution [0] for session recordings alongside basics like DAU and revenue. [0] https://ww
502.
by XCSme 8mo ago
Unique visitor counts are definitely useful, and you dont need Google Analytics just for that. A self-hosted solution [0] can give you the basic numbers and trends you need without using third-party tracking. [0] https://www.uxwi
503.
by XCSme 8mo ago
This is exactly why so many people are thinking about self-hosting everything. Analytics is another big point of failure and a potential privacy risk that often gets overlooked. Relying on third-party tools means any outage, policy change,
504.
by XCSme 8mo ago
I’m the founder of UXWizz, and a lot of founders I know use it when they want something self-hosted so they fully own their customer data. It’s a lot cleaner than a raw BI tool and works well for simple client dashboards or “stats-as-a-serv
505.
by XCSme 8mo ago
Your landing page is cool. Why would I need WebSockets? They do have to connect to something, so I assume I would already have a server. Or is it for p2p connections?
506.
by XCSme 8mo ago
All AI landing pages are so confusing, maybe because they are AI-built as well...
507.
by XCSme 8mo ago
This looks really cool! One thing that can help a lot early on is seeing how people actually use the landing page, not just if they click. Stuff like session replays or heatmaps can show where people get stuck[0], what grabs attention, and
508.
by XCSme 8mo ago
Yeah, once you start self-hosting your code, it’s kind of nice having control over everything. Makes you think about moving other stuff, like analytics, to something self-hosted too.
509.
by XCSme 8mo ago
Hey, that's cool. I know the Show HN rules only allow sharing products that are free to install or provide a live-demo with no sign-up required, is there any demo dashboard available?
510.
by XCSme 8mo ago
Many developers want to understand user behavior without the heavy baggage of ad-tech tracking. What do you think of self-hosted tools (like UXWizz) to track qualitative and quantitative data, but without having the problem of web-wide surv

More ›