Gebbia wants to change the retirement process from its current paper-based system to “an online digital process that will take a few days at most.”
This can only go wrong, like it always does when some fancy big idealist tries to take a limping but still working system and overhaul it to a full digital "seamless and easy" procedure.
I have no doubt the current process is inefficient, because its the federal government, but these people retiring are getting a lot of benefits. It should take a moment to make sure they qualify properly and are getting only what they should be getting. They could probably cut a lot of the inefficiency while still keeping it fully hard copy and reliable.
I like a lot of what DOGE is doing, but we should also be weary of reaching the other extreme, where we lose site of doing things fully properly just to make it quick and minimalist, or just part of Musk's futurist all digital utopian ideals.
What, you think going through 3 months of paid training and 2 months of afterhours training in order to file specific documents and send them around the country (paper only, no email), which then are re-assembled a salt mine (again paper-only, no emailing them) to be brought down a once-a-week elevator which is then, a week later, verified by another person, at which point, another week later, another third party comes in and files just that one paper set (other retirees need to wait) could POSSIBLY have ANY inefficiencies to it?
I remember working for the 2010 census and they were very excited to tell us that we were getting laptops with 56k modems to take home. But you needed a landline to update anything and it had to be the old style landline that didn't connect to a modem.
Thank you for summing it up better then I could have, another thing to mention is the long trip by buggy down in the mine which I think is about half an hour drive to reach the spot where new documents are added? I could be wrong on that though.
It genuinely is, fundamentally ridiculous.
Like, I do prefer backup paper copies of things, but...I do not think this system should be the primary system...it should be done as a backup and thus having a lag between "new data dumps" would be fine. It just doesnt work for something that requires more spontaneous and more immediate responses though, so it should never be a primary system.
It says in the article it takes around two months, which is pretty bad.
There is no doubt a lot of improvement to be made to it, but going full digital and focusing on making it super duper fast is almost guaranteed to bring a whole host of other problems by going to the opposite extreme.
I've worked enough random jobs where we had a shitty, but workable system for some random task, and a new Exec with bright ideas came in and tried to fully automate and digitize it, only for that to be riddled with so many failures and problems we end up taking triple the time, losing mass productivity/profit due to failures, and just going back to the old system anyway. While said Exec cashes out and leaves us all holding the bag of fixing the mess.
Which is what I foresee happening based on what this guy is putting his focus on, not improving the system but making it shiny and futuristic and just like Apple.
I definitely understand where you are coming from, but as another poster summed up the current method just flat out needs to be changed, it just is untenable how ridiculous it is.
And...frankly if there is anyone that actually has a proper proven track record of actually fixing shit software side, I'd say Elon Musk has a decent one. Easy example would be twitter/X which runs remarkably well with less than half the staff it did. Hell, I think its had fewer outages (at least, per year) compared to before Elon acquired it.
That and Elon seems to know how to properly manage data engineers and guide them properly, so frankly if anyone can do it, its probably Elon.
The only way it could possible work out well at least at first is if the organization already had a real plan they wanted to do but just couldn't make the administrators do it.
It sounds like a lot of departments already knew what was wrong and how to fix it, but they'd have to fire thousands and couldn't.
This can only go wrong, like it always does when some fancy big idealist tries to take a limping but still working system and overhaul it to a full digital "seamless and easy" procedure.
I have no doubt the current process is inefficient, because its the federal government, but these people retiring are getting a lot of benefits. It should take a moment to make sure they qualify properly and are getting only what they should be getting. They could probably cut a lot of the inefficiency while still keeping it fully hard copy and reliable.
I like a lot of what DOGE is doing, but we should also be weary of reaching the other extreme, where we lose site of doing things fully properly just to make it quick and minimalist, or just part of Musk's futurist all digital utopian ideals.
If memory serves, the current method takes months so honestly I'd be surprised if they found a way to make it worse.
Seriously the current retirement process is kind of comical.
What, you think going through 3 months of paid training and 2 months of afterhours training in order to file specific documents and send them around the country (paper only, no email), which then are re-assembled a salt mine (again paper-only, no emailing them) to be brought down a once-a-week elevator which is then, a week later, verified by another person, at which point, another week later, another third party comes in and files just that one paper set (other retirees need to wait) could POSSIBLY have ANY inefficiencies to it?
I remember working for the 2010 census and they were very excited to tell us that we were getting laptops with 56k modems to take home. But you needed a landline to update anything and it had to be the old style landline that didn't connect to a modem.
Thank you for summing it up better then I could have, another thing to mention is the long trip by buggy down in the mine which I think is about half an hour drive to reach the spot where new documents are added? I could be wrong on that though.
It genuinely is, fundamentally ridiculous.
Like, I do prefer backup paper copies of things, but...I do not think this system should be the primary system...it should be done as a backup and thus having a lag between "new data dumps" would be fine. It just doesnt work for something that requires more spontaneous and more immediate responses though, so it should never be a primary system.
It says in the article it takes around two months, which is pretty bad.
There is no doubt a lot of improvement to be made to it, but going full digital and focusing on making it super duper fast is almost guaranteed to bring a whole host of other problems by going to the opposite extreme.
I've worked enough random jobs where we had a shitty, but workable system for some random task, and a new Exec with bright ideas came in and tried to fully automate and digitize it, only for that to be riddled with so many failures and problems we end up taking triple the time, losing mass productivity/profit due to failures, and just going back to the old system anyway. While said Exec cashes out and leaves us all holding the bag of fixing the mess.
Which is what I foresee happening based on what this guy is putting his focus on, not improving the system but making it shiny and futuristic and just like Apple.
I definitely understand where you are coming from, but as another poster summed up the current method just flat out needs to be changed, it just is untenable how ridiculous it is.
And...frankly if there is anyone that actually has a proper proven track record of actually fixing shit software side, I'd say Elon Musk has a decent one. Easy example would be twitter/X which runs remarkably well with less than half the staff it did. Hell, I think its had fewer outages (at least, per year) compared to before Elon acquired it.
That and Elon seems to know how to properly manage data engineers and guide them properly, so frankly if anyone can do it, its probably Elon.
The only way it could possible work out well at least at first is if the organization already had a real plan they wanted to do but just couldn't make the administrators do it.
It sounds like a lot of departments already knew what was wrong and how to fix it, but they'd have to fire thousands and couldn't.