All Activity

--I wouldn't call it "disdain", per se. But definitely reservations. One should always understand "how" things work. AI is at a sort of 'intellectual event horizon' beyond which it cannot be fathomed - not by 98% of the using population. This is not an absolute either. There is a crisis of social mores, and breakdown of common morality. Example, that that used to keep mass shootings ( for example) down to one or two every 10 years going back to the 1920s. All but just a very small tiny fraction of beyond-fringe psychosis were at least sufficiently kept in check. Now? one ever 6 months, a number that exceeds the "very small fraction" number. Something is motivating enough to indoctrinate merely untoward ideologies into specific actions. And if anyone has a modicum of understanding about sociology and history, they'll see that the most fantastic force to ever have impacted humanity, during and immediately preceding that uprising, wasn't just fire or the wheel. It has been this unchecked innovation --> advent of technologies so profoundly capable of unilaterally either helping, or afflicting at a species level, that the only fair way to define this era is truly a techno-sociological experiment at an evolutionary scale - one that 90some% of the population density will lack the capacity to even be aware of what it is they are being subjected to ... Can't end well playing with Cosmos' gun. Failure to understand and commit to the virtuosity of understanding how something works, that same something that one is allowing to guide them? That is a problem when people intertwine with it in blind presumptive faith. From one end of the spectrum of miss-use, to the other end of becoming co-dependent, then having it fail: Where does that leave us?

I've been raking the edges to minimize the icicles. It's a pita but I don't mind, it's al part of the clean up routine now.

No disagreement there, but (and maybe I am flat out wrong on this) we can't teach AI something that we don't know ourselves. For example, when it comes to physics and mathematics, AI isn't going to teach us or give us a better understanding of how atmospheric physics works and how these processes behave and evolve. Let's look at thunderstorms, for example. There are certain processes which occur during a thunderstorms life cycle that we know happen, however, we don't fully know why certain processes happen the way they do or what is the leading contributor. Tornadogenesis is one...we know the ingredients needed for tornadoes, we know how tornadoes form, but we don't know fully understand why some supercells (which look tornadic based on visual features/radar features) produce tornadoes and others don't...AI isn't going to solve something like that. AI isn't going to tell us this because we don't understand it ourselves and we don't have all the necessary data and measurements to be able to do so. But with the phone example, its learning faster because it has a basis to go on...it understands something because that something is known. There are many meteorological processes which we know exist and understand their existence, but don't fully know the why/how. AI will be a major help though in calculating mathematical/physics calculations much more quickly which will hopefully get us faster model output in the future.

I think you’re right about this North Hills. Not seeing anything definitive to indicate blocking and we definitely don’t have a good western ridge. We will have to be patient. Perhaps we will start to see things turn around on by late December as negative height anomalies begin to wane in the eastern Pacific and positive heights begin to build east of Greenland and possibly push west. .

I love building a base of my own resources....I sometimes draw upon them when the situation warrants. It's also nice when some douche misquotes me, as I simply copy/paste my actual thoughts in short order.

The snow drought continues in WNC. And the dry conditions will continue through the end of the month. Its almost a guarantee December will be way below average rain for most of the state.

Here is another map at 6 days.

Its the EPS and I can show 200 hour maps that show the same. Anywhere from 6-16 days look at or above average according to the EPS.

Paul, I think your approach narrows the possibilities. I mean isn’t AI inherently “better computing” considering the concept is deep learning? I think AI can be tweaked when we see where it needs to be quickly. As a crude example, my phone quickly learns how I use “muthufukka”. Phones five years ago took much longer and my next phone should even learn it faster.

Central Park was updated to 2.9” lol

17.9 this am.

ONI should def. peak in OND period.

It’s very concerning with all the money now trading on predict markets where CPK measurements are the benchmark

It wouldn’t be the first low snowfall measurement there. I have pointed this out in the past with multiple other posters here. The actual measurement was probably closer to the 4.0” at Newark. Thanks for posting the video. It may have been Sam Champion last year showing closer to 1” too low in one of the events where he used his own measurement.

As @wncsnow is saying, I see nothing muted about this “torch”. To me a week of above normal for the eastern half of the us is a torch. Maybe our temps get muted somewhat bc of a wedge but synoptically the cold is gone from the east for an extended period so this is the dreaded shut the blinds pattern until maybe the first week in January. I don’t see a hell ridge from the SER but if you want winter weather (everyone posting here) near normal isn’t going to cut it and there aren’t any threats as far as we can look out at this point. We’re punting 2+ weeks that’s a given.

Kind of explained it in a post above but I'll add more. We should be focusing on the resources needed to better what we currently have instead of just adding more tools to the toolbox. Now, if AI will be used to better the initialization/parameterization process, that would be amazing. But at the end of the day we still need better computing technology (which we have, we just need it within the field). Quantum computing is going to go a sizable way I think here. Parameterize better and improve initialization, those two alone will go a great way in forecast model accuracy and hopefully reduce inconsistency. If AI models are just going to add to the list of potential outcomes and increase uncertainty, then what good is it?

No, you’re not playing right, you need to constantly be dissatisfied and fretting about the next storm threat.

Got down to 12.7 here in Pleasant Gardens. I had a couple wind gusts yesterday exceed 25mph.

You're posting a plus 300 hour map bro.

Yeah I don’t understand it either. It’s a new tool that should improve over time. Gotta keep pushing forward with progress. Some paths are deadends, but some become the main artery.

Thanks for the explanation. In theory though, they should end up more skillful than what we have now..so it can't be much worse than current day?

Frozen bare ground. How nice.

Hopes and dreams of snow never did anyone any good. Reality is a bitch, and right now the bitch is having her way driving weenies nuts.

@WxWatcher007 why i dont have gutters, old house with crap insulation in attic = death icicles.

Essentially the best chess AI's are self taught so that may be the best approach to take with AI weather models as opposed to teaching the model(s)