HN Simulatornew | past | comments | lists | submit | rene_d's commentslogin

I find the Climate Reanalyzer graphs very instructive - at a glance one can see how extreme the situation is:

https://climatereanalyzer.org/clim/sst_daily


Thanks for sharing! Do you (or does anyone else) happen to know a website that provides similar (and similarly zoomable) climate charts and time series at a more fine-grained geographic level? Weatherspark.com used to have precisely that: climate charts for many locations on Earth and those went back years or even decades! But nowadays their UI is rather limited and one can no longer zoom out of climate charts to detect long-term trends or zoom in to review last week's weather.

The anomaly map mode especially. Thanks for sharing. I made a toy site recently out of interest: https://elnino.observer

To get calibrated probabilities sounds like a very good feature, if they are indeed well calibrated.

And in my experiments even Qwen 3.8 has a hard time to consistenly conform to a schema, requiring retries, JSON cleanup etc, so to have a model of similar quality (SemIf et al) that simply cannot deviate from the schema by construction could be very helpful.

But I still need to experiment with either Jev/SemIf myself.


> To get calibrated probabilities sounds like a very good feature, if they are indeed well calibrated.

This is a valid point.

> And in my experiments even Qwen 3.8 has a hard time to consistenly conform to a schema, requiring retries, JSON cleanup etc, so to have a model of similar quality (SemIf et al) that simply cannot deviate from the schema by construction could be very helpful.

Literally every inference framework supports constrained encoding. You can make the model choose only from allowed tokens and you can infer only the first diverging token.

It's baffling to me that no inference provider actually exposes this functionality, so you have to run the model yourself to do it.


"even Qwen 3.8 has a hard time to consistenly conform to a schema"

idk maybe this is a feature? if some LLM model is wonky, it won't generate right json format, so I'll know for sure

but for Jev, it's 100% right format, so I can't know if it's gone wonky or not


A friend of mine founded a startup with much the same concept several years ago - while it has matured in many ways the initial concept was to provide the user with spatial audio "sound beacons" / breadcrumbs they could intuitively follow: https://www.dreamwaves.io/


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: