The thread keeps circling around the politics, but almost nobody has dug into what actually goes on in the NWS tornado warning pipeline.
It's worth being specific: the National Weather Service operates some of the most robust automation and radar ingest pipelines on Earth, but the final go/no-go warning call is almost always human—often a single overnight forecaster on a console, monitoring a swath of counties. Automation (e.g., Warn-on-Forecast guidance) can surface threats, but the NWS intentionally doesn't have an 'auto-warn' button for tornadoes, because of the asymmetry of false positives (blow credibility, cost lives in the long run).
Budget cuts reduce redundancy and experience in those overnight shifts. When you have only one person monitoring instead of a team of two or three, you get decision fatigue and coverage holes, especially during clustered, multi-cell outbreaks. We've seen near-misses in the past, and every pro-meteorologist I know says they're playing defense against process errors, not just technology failures.
Before we point fingers or blame 'technology/automation' shortfalls, let's quantify the concrete bottleneck: skilled human decision-makers are the limiting reagent; machine learning warning aids are still years away from majority trust.
>Before we point fingers or blame 'technology/automation' shortfalls, let's quantify the concrete bottleneck: skilled human decision-makers are the limiting reagent
All the automation in the world with useless without a human guide to either transform production into a useful product, or useful knowledge to heed. That's why this act of trying to remove human labor is asinine. Even skilled human can't always get the right readings, so expecting a robot to do it all at this stage is just selling snake oil.
Actually, I'll go a step further - in the long run, we probably won't need human forecasters at all.
The current human-in-the-loop model exists largely because our technology hasn't been good enough yet, not because there's something inherently special about human judgment in this context. Weather prediction is fundamentally a pattern recognition problem. Pattern analysis at scale is exactly what computers do better than us.
yeah why cannot that guy sit in california or new york in a normal time zone? not like there are tornadoes in every state, its so silly to keep a person at night in an office when weather is good
There's only a 6 hour different between the East coast and Hawaii. You can't entirely avoid a night shift, so you might as well have them all work from the same location.
Fortunately the Kentucky NWS office in Jackson was fully staffed during the recent events. It's still not staffed 24/7, but at least they bring in people when things inclement weather is occuring.
They may have been been "fully staffed" according to the "new" "fully staffed" number which could be not enough to perform all the duties and responsibilities reliably and without disruptions. I've been on teams that ought to have 5-10 people minimally wearing all the hats but are only by budgeted for and allow 3 by upper management. Those teams, according to the decision makers are fully staffed, even though they are inadequate.
Edit: and during death marches those can be worked overtime on weekends to make a deadline. They're "fully staffed" and work for the short-term until burnout and turnover begins. Then those projects are always behind.
This seems entirely plausible, but I'm not sure this article successfully makes this connection with direct evidence. Has anyone seen anything on this with better evidence?
NYT mentioned NWS staffing shortage, but did not say this was connected to body count:
"The office is also one of several left without an overnight forecaster, but on Friday, it stayed open and was sufficiently staffed for the night, issuing 11 tornado warnings. It was “all hands on deck,” Mr. Fahy said."
That's because no one can make that conclusion definitively yet. They want your brain to assume that connection. Conspiracy theorists are the kings at this psychological trick.
Thanks! The information here is why I think it's extremely plausible that staffing shortages were disastrous here. But it also doesn't make the direct link. I think it's probably just too soon to know whether all the warnings went out as they should.
But another problem with our current government is that I'm skeptical any investigation to answer those questions will ever happen.
Sadly anyone with a a half of a brain saw this coming. To add to this, FEMA cuts probably means these poor people will live through what Puerto Rico did when the island was wiped out years ago by that hurricane :(
I wish they could sue Trump and Musk personally for making dumb decisions.
Part of me doesn’t feel too much sympathy. The majority of Americans voted for Trump. Even more so in the conservative states that are more likely to be impacted by there events.
Trump campaigned on cutting government services.
Everyone is okay with cutting a public service (at the expense of others) until they need that particular service
---
To clarify, I'm not cheering on this disaster or hoping that those who voted for Trump "get what they deserve"
In states like Kentucky a d Oklahoma? Think again. These places are overwhelmingly trumpy. The nation as a whole barely voted him in, but in many red states his numbers are very high. My home county in Tx voted 80% Trump.
Politics aside, it’s odd how often the entire debate misses the real bottleneck: assigning blame doesn’t restore operational capacity or re-architect the warning pipeline. If the system depends on 24/7, highly skilled human decision-makers, and you cut those positions, the outcome is predictable—slow, brittle responses.
ANY critical pipeline that can be broken by one missing seat is overdue for technical reinforcement
Unfortunately, a lot of liberals live in these states too. This entire country would be blue if we didn't have the electoral college.
But if you're in the mood for not feeling sympathy, because most cities have heat domes that tend to "push" storms away from them, storms tend to leave the more-liberal cities alone and instead wreak havoc through more rural towns.
Welcome to 2025, where we gather the family around a glowing laptop watching Ryan Hall Y'all and his YouTube channel telling us when it's time to take shelter.
Readit News
It's worth being specific: the National Weather Service operates some of the most robust automation and radar ingest pipelines on Earth, but the final go/no-go warning call is almost always human—often a single overnight forecaster on a console, monitoring a swath of counties. Automation (e.g., Warn-on-Forecast guidance) can surface threats, but the NWS intentionally doesn't have an 'auto-warn' button for tornadoes, because of the asymmetry of false positives (blow credibility, cost lives in the long run).
Budget cuts reduce redundancy and experience in those overnight shifts. When you have only one person monitoring instead of a team of two or three, you get decision fatigue and coverage holes, especially during clustered, multi-cell outbreaks. We've seen near-misses in the past, and every pro-meteorologist I know says they're playing defense against process errors, not just technology failures.
Before we point fingers or blame 'technology/automation' shortfalls, let's quantify the concrete bottleneck: skilled human decision-makers are the limiting reagent; machine learning warning aids are still years away from majority trust.
Deleted Comment
All the automation in the world with useless without a human guide to either transform production into a useful product, or useful knowledge to heed. That's why this act of trying to remove human labor is asinine. Even skilled human can't always get the right readings, so expecting a robot to do it all at this stage is just selling snake oil.
The current human-in-the-loop model exists largely because our technology hasn't been good enough yet, not because there's something inherently special about human judgment in this context. Weather prediction is fundamentally a pattern recognition problem. Pattern analysis at scale is exactly what computers do better than us.
Perhaps someone could apply to YC with this idea. There is one YC startup doing this already: https://www.ycombinator.com/companies/atmo
https://www.weku.org/the-commonwealth/2025-05-17/kentucky-nw...
Edit: and during death marches those can be worked overtime on weekends to make a deadline. They're "fully staffed" and work for the short-term until burnout and turnover begins. Then those projects are always behind.
But another problem with our current government is that I'm skeptical any investigation to answer those questions will ever happen.
I wish they could sue Trump and Musk personally for making dumb decisions.
Why didn't a well-funded FEMA prevent what happened in Puerto Rico?
Deleted Comment
You do not need a radio license as things can be called in by phone, too.
https://www.weather.gov/skywarn/
Trump campaigned on cutting government services.
Everyone is okay with cutting a public service (at the expense of others) until they need that particular service
---
To clarify, I'm not cheering on this disaster or hoping that those who voted for Trump "get what they deserve"
ANY critical pipeline that can be broken by one missing seat is overdue for technical reinforcement
Deleted Comment
Dead Comment
But if you're in the mood for not feeling sympathy, because most cities have heat domes that tend to "push" storms away from them, storms tend to leave the more-liberal cities alone and instead wreak havoc through more rural towns.
Dead Comment
Dead Comment