From Wired:
The so-called Department of Government Efficiency (DOGE) is starting to put together a team to migrate the Social Security Administration’s (SSA) computer systems entirely off one of its oldest programming languages in a matter of months, potentially putting the integrity of the system—and the benefits on which tens of millions of Americans rely—at risk.
The project is being organized by Elon Musk lieutenant Steve Davis, multiple sources who were not given permission to talk to the media tell WIRED, and aims to migrate all SSA systems off COBOL, one of the first common business-oriented programming languages, and onto a more modern replacement like Java within a scheduled tight timeframe of a few months. ...
SSA’s core “logic” is also written largely in COBOL. This is the code that issues social security numbers, manages payments, and even calculates the total amount beneficiaries should receive for different services, a former senior SSA technologist who worked in the office of the chief information officer says. Even minor changes could result in cascading failures across programs.
“If you weren't worried about a whole bunch of people not getting benefits or getting the wrong benefits, or getting the wrong entitlements, or having to wait ages, then sure go ahead,” says Dan Hon, principal of Very Little Gravitas, a technology strategy consultancy that helps government modernize services, about completing such a migration in a short timeframe.
You may recall that Frank Bisignano testified at his confirmation hearing that COBOL was still widely used in business and that its presence at Social Security was nothing to be too concerned about.
29 comments:
Brilliant idea and long overdue. Imagine what a gift it is to Social Security to have the greatest minds in tech build it a brand new system. Inspiring.
Big Ballz to the rescue!
Yeah, the brilliant genius minds at work, no one ever thought to try to migrate off the mainframe and COBOL. We only try ever couple years when we think we'll have the appropriate support, only to start, realize it's a massive undertaking, ask for proper funding to fuel efforts, get denied and the cycle starts over again.
Sadly, no satire codes so 2:24 is a troll. Hey, let's pretend it's a plane at 30,000 feet with a ton of passengers and these "greatest minds in tech" wanna rebuild the flight systems while it is in the air. Oh, and they aren't actually on the plane, so if they break it, there is no personal impact on them.
Yes. Have a bunch of kids with no COBOL experience redo the code. Gosh, what could go wrong?????
Just run the new code in parallel with the production COBOL for a year and see if/what the differences are between the two. Adjust new code whenever necessary while running in parallel until results are stable and exactly same. While parallel testing ongoing, new coding will not impact production code.
Some of our systems have issues in the first place with workarounds and theyve been around for years, how could you do this in months even with AI? Also what if AI hallucinates and gives code thats bad feedback.
Imagine thinking the greatest minds in tech were the 30 douche-nozzles that Elon assembled to rape & pillage the government.
Yeah, break things into logical and manageable pieces, come up with validation cases for testing and debugging, work out how to parallel test efficiently and establish a protocol to see that the inputs create the proper outputs and production products. Test that those products work in whatever system is taking them as inputs. That's just changing the codebase, no change to functionality. Changing functionality or design is a lot different and harder.
Either way, shouldn't take more than a few months. Not.
How will they pay for it? Or...is Musk donating the services?
We’ve been dual-processing cases at OHO trying to migrate to DCPS for years now. We are 2 years past the projected deadline for elimination of the one we solely used for 20 years. It is so time consuming. A new system, with fewer staff than ever, sounds very efficient.
This will not end well. Prove me wrong.
There’s a whole book about this general issue titled Recoding America, written by someone with experience at the highest levels of both Silicon Valley and the government. The gist is that it’s not as easy as Silicon Valley thinks it is, but it’s more feasible than the government thinks it is.
https://www.recodingamerica.us/
Watch out for a rapid unscheduled disassembly of SSA systems.
Rumor has it NSD is getting dissolved but where’s Charles? At the bar enjoying happy hour, methinks!
Congressman Beyer called the 800#...told the wait time > 120 minutes and he was disconnected
https://bsky.app/profile/beyer.house.gov/post/3llhos643l22f
What could go wrong? Anything and everything. And it will go wrong, as planned.
The Trump/Doge intent appears to be to bleed and break the system so irretrievably that it provides a basis for doling out expensive government contracts to the private sector to "fix" Social Security, i.e., privatization. Impact will fall hardest on the cult members who attended the rallies and unwittingly helped facilitate this.
And how much does that cost?
The payments will be disrupted. Get ready for all hell to break loose because it will happen thanks to Sleazy E!
@2:24 PM, March 28, 2025 Where is the money coming from to do this? The agency can't even accelerate SMS or digital comms, or get documents in digitally without having to rekey everything in downstream systems. It's already bad enough that OCIO has dragged benefit systems into three platforms and technicians can't do the job -- perhaps these smart people can put their plan in the public sphere or on DOGE's website or on Twitter, so the world can see exactly how they plan to collapse it all.
The real problem has been leadership. No real vision or roadmap showing where we are and where the agency wants to go. This will need an enterprise wide plan to execute, implement and migrate. It has nothing to do with legacy systems and all to do with leadership. So many executives have been managers and not good ones at that level. Making small tactical moves that have caused the current problems. IT Mod lacked vision as to how to modernize the disability determination process. It was piecemeal favorite projects of executives to get their bonuses. SSA has lacked a real visionary leader for a long time. Leathers is no silver bullet here, but a leader with a good plan and Target goals can get us there. There are a number of great efforts going on but all increment parts of the process in reducing case processing time and cost here and there.
@11:26am Well put, as "Leathers" I presume is a typo for Leadership.
Well, I think DCPS cost over $300 million and failed, twice, so doing the same thing again and expecting different results doesn't seem logical.
With a career of 40+ years in software engineering, I've learned (and forgotten) Cobol at least twice. It's a simple language. So simple, you can even write a program to translate Cobol programs to one of those "modern" languages.
Great! But what have you accomplished? It's fun and popular to blame the old, but programmed tasks work or fail just as well in any programming language. As hinted in other comments, some of this has been previously tried, but it appears it did not go well. I see no explanation, never mind transparency, what is different.
I suggest he call a local office number and see how quickly he gets through. That is after the two minute inane preamble before you even get to start pushing buttons to get where you want to go, even if you know an extension, or just to wait for a call back that never comes.
Updates: DOgE rebuild SSA databases
https://www.wired.com/story/doge-rebuild-social-security-administration-cobol-benefits/
Elon Musk says his AI company and X are merging
https://www.washingtonpost.com/technology/2025/03/28/xai-acquires-x-elon-musk/
Both events are tied to Musk using SSA data to train AI Grok to create an Twitter "everything" app that controls your social media, banking, shopping, financials and is a locator device.
Crazy talk
Interesting- Google AI says that it can take 5 to 10 years to migrate complex COBOL systems.
Post a Comment