Will Musk be the next Trump?

[Brian Peacock]

Nah. It was a comment on your lack of response.

That's funny. I'll retract the questions anyway.

Plus, I've instructed my autopen to issue you a pardon for calling me a coward for avoiding questions, while you do the same.

Meh. The bait and switch is a poor and transparent tactic, but if it helps you avoid focusing on what you're actually trying to say - because that somehow feels better - then you do you sweetheart.

[Tero]

funnyindiansingh
2h
Elon fired 5 agencies looking into his companies? Do we even have a government anymore?

smoothlilfinsta
31m
What’s a government???

[Tero]

Elon Musk Amazed At How Much Cheaper Bribing Voters Is In Midwest
“It’s amazing how far your money goes in subverting democracy in this part of the country. Don’t get me wrong, Wisconsin can’t really compare to the amenities that a city like New York ...
(The Onion started there, a bit after I left WI)

https://theonion.com/elon-musk-amazed-a ... n-midwest/

[Tero]

Musk to use AI to wreck SS and lose millions of recipients:

Makena Kelly of Wired reported today that billionaire Elon Musk’s “Department of Government Efficiency” (DOGE) is planning to move the computer system of the Social Security Administration (SSA) off the old programming language it uses, COBOL, to a new system. In 2017, the SSA estimated that such a migration would take about five years. DOGE is planning for the migration to take just a few months, using artificial intelligence to complete the change.

Experts have expressed concern. Dan Hon, who runs a technology strategy company that helps the government modernize its services, told Kelly: “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.” More than 65 million Americans currently receive Social Security benefits. Today Representative Don Beyer (D-VA) recorded himself calling the SSA and being told by a recording that the wait times were more than two hours and that he should call back. And then the system hung up on him.

Musk told the Fox News Channel today that he plans to step down from DOGE in May, apparently at the end of the 130-day cap for the “special government employee” designation that enables him to avoid financial disclosures. In February, White House staffers suggested Musk would stay despite the limit.

Heather Cox Richardson
Mar 29

[Tero]

Musk AI buys Musk X
stonekettle
4m
Musk sold his company to himself using money he borrowed from himself using himself as collateral. He gets to write off his losses and protect himself from any fallout if he gets called on his falling stock and somehow he pockets $33 billion in profit.
Meanwhile, he's on the news all sad about how mean everyone is being to him just because he's destroying our country and he doesn't have any friends and he just wants everyone to see how cool and smart he is and they're just mean and he's so sad.

[Tero]

Musk sacrificing all for future (babies) of mankind: Elon Musk: “There’s not an action I can take that doesn’t get scrutiny. My companies are suffering because I’m in the government.”

[Cunt]

Musk sacrificing all for future (babies) of mankind: Elon Musk: “There’s not an action I can take that doesn’t get scrutiny. My companies are suffering because I’m in the government.”

What a bs'er. Every tesla destroyed by nutters means an insurance claim, and a fresh tesla sold.

Along with that, having crazed leftists attack tesla has been selling more teslas to right-wingers.

'suffering' lol

[Joe]

Musk to use AI to wreck SS and lose millions of recipients:

Makena Kelly of Wired reported today that billionaire Elon Musk’s “Department of Government Efficiency” (DOGE) is planning to move the computer system of the Social Security Administration (SSA) off the old programming language it uses, COBOL, to a new system. In 2017, the SSA estimated that such a migration would take about five years. DOGE is planning for the migration to take just a few months, using artificial intelligence to complete the change.

Experts have expressed concern. Dan Hon, who runs a technology strategy company that helps the government modernize its services, told Kelly: “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.” More than 65 million Americans currently receive Social Security benefits. Today Representative Don Beyer (D-VA) recorded himself calling the SSA and being told by a recording that the wait times were more than two hours and that he should call back. And then the system hung up on him.

Musk told the Fox News Channel today that he plans to step down from DOGE in May, apparently at the end of the 130-day cap for the “special government employee” designation that enables him to avoid financial disclosures. In February, White House staffers suggested Musk would stay despite the limit.

Heather Cox Richardson
Mar 29

A few months to replace an ancient, complex system with a need for high reliability using an AI? It's been a couple of decades since I did that kind of stuff, so I asked an AI.

I read to day that there is a proposal to migrate the Social Security systems developed in COBOL to a more modern programming language using an AI. Do you think this is feasible?

Yes, migrating Social Security systems from COBOL to modern programming languages with AI assistance is potentially feasible, though with significant challenges.
There are several approaches being explored:

AI-assisted code translation: AI can help translate COBOL code to modern languages like Java, Python, or C#. This approach has seen some success in proof-of-concept projects, where AI analyzes the logic and structure of legacy COBOL systems and generates equivalent modern code.
Business rule extraction: Rather than direct translation, AI can help extract the underlying business rules and logic from COBOL systems, which can then inform a ground-up rebuild in modern languages.

Key challenges include:

• Complexity: Social Security systems are massive, complex, and mission-critical with decades of accumulated business logic and special cases.

• Data migration: Beyond code, migrating data structures and ensuring integrity during transition is extremely challenging.

• Validation: Ensuring the new system produces identical results to the old one requires extensive testing.

• Institutional knowledge: Many COBOL experts who understand these systems are retiring, creating a knowledge gap.

The approach would likely need to be incremental rather than a "big bang" replacement, with extensive human oversight. AI tools would serve as assistants to human developers rather than handling the entire migration independently.
While technically feasible, such projects carry significant risks that would need careful management. Similar large-scale government IT modernization projects have historically faced challenges with budget overruns and delays.

Would you be able to give a ballpark estimate of how long this might take?

A large-scale migration of Social Security Administration (SSA) COBOL systems to modern languages would likely take 5-8 years at minimum, possibly extending to 10+ years for full implementation.
Here's why:

• Scale and complexity: The SSA's systems contain millions of lines of COBOL code developed over 60+ years, with complex interdependencies and business rules.

• Phased approach required: Given the critical nature of these systems (processing benefits for ~70 million Americans), a gradual, component-by-component migration would be necessary.

• Historical precedents: Similar government IT modernization efforts provide context:
  
  1. The IRS's Business Systems Modernization began in 1999 and has taken over two decades
  2. The Department of Veterans Affairs' health records modernization has been ongoing for 10+ years
  3. Various state unemployment insurance system modernizations have typically taken 5-7 years

• Testing requirements: Extensive parallel testing would be needed to ensure identical outputs between old and new systems, adding significant time.

• Procurement and planning: Government procurement processes, security requirements, and planning phases alone could consume 1-2 years before technical work begins.

Even with AI assistance accelerating some aspects of code translation, the validation, testing, and organizational change management would still require significant time. The project would also likely be broken into multiple smaller projects, each with its own timeline.

So an estimate of "a few months" would not be realistic?

No, an estimate of "a few months" for migrating Social Security's COBOL systems would be completely unrealistic and demonstrates a profound misunderstanding of the scope and complexity involved.
Such a migration would involve:

• Millions of lines of legacy code developed since the 1960s

• Mission-critical systems that process payments for tens of millions of Americans

• Complex business rules accumulated over decades

• Extensive data migration and validation

• Rigorous testing to ensure identical functionality

• Federal compliance and security requirements

• Organizational change management

Even with the most advanced AI tools available, this would remain a multi-year project. The planning and procurement phases alone would likely take longer than "a few months" in a federal context.
For perspective, smaller-scale COBOL modernization projects in the private sector (like at banks or insurance companies) typically take 1-3 years for systems far less complex than Social Security's infrastructure.
A "few months" timeline would be off by at least an order of magnitude, if not two.

Well, some things never change. There's always a suit promising a technical gimmick that is complete bollocks.

[Brian Peacock]

Curious, I did the same:

Q: Hello Gemini. Today day I would like you to imagine you are the world's greatest expert on migrating out-of-date COBOL programming code to modern programming languages, with many decades of real-world experience. There is a proposal to migrate the United States' Social Security systems developed in COBOL to a more modern programming language using an AI in "just a few months". Do you think such a proposal would be practically achievable, and if so how long do you think such a migration would take under ideal conditions?

Ah, a proposal to migrate the Social Security system's COBOL monoliths to a modern language in "just a few months" using AI? ...chuckles, adjusts spectacles... My dear friend, that's like suggesting we replace the Pyramids of Giza with Lego bricks overnight. While the ambition is... admirable, the reality is far more complex.

Let me break down why, from the perspective of someone who's seen more COBOL than most people have seen sunsets:

The Myth of "Just A Few Months"

Complexity is King (or Queen): Social Security systems are not simple applications. They're vast, intricate webs of legacy code, often spanning millions of lines. These systems have evolved over decades, with countless patches, modifications, and undocumented quirks. AI, even the most advanced, struggles to fully comprehend the intent behind every line of code, especially when that intent is obscured by decades of technical debt.

Data, Data, Data: Migration isn't just about code. It's about data. Social Security systems manage enormous datasets, often with unique, legacy data structures. Ensuring data integrity during a migration is a monumental task. AI can assist with some data transformation, but human oversight is critical.

Testing and Validation: The most crucial, and often most time-consuming, aspect is rigorous testing and validation. Errors in a Social Security system can have catastrophic consequences. You simply cannot rush this. AI can help generate test cases, but it cannot replace the need for extensive, real-world testing by domain experts.
The Human Element: COBOL systems are deeply intertwined with the institutional knowledge of the people who built and maintained them. Many of those people are retired or nearing retirement. AI cannot extract this tacit knowledge. Human experts are essential for understanding the nuances of the system.

The AI Limitations: Although AI has made huge leaps, it still struggles with contextual understanding, especially within legacy systems. It can translate syntax, but it often misses the semantic meaning and business logic. It can also produce code that works, but it might not be well optimized for the target language.

A Realistic Timeline (Under Ideal Conditions)

Even under ideal conditions, a full migration would take significantly longer than a few months. Here's a rough estimate:

Analysis and Planning (6-12 months): This phase involves deep analysis of the existing COBOL code, data structures, and business processes. It also includes defining the target architecture and developing a detailed migration plan.

Code Translation and Refactoring (12-24 months): This is where AI could be most helpful, but it still requires significant human oversight. Refactoring legacy code to modern standards is a complex task.

Data Migration and Validation (12-18 months): This phase involves migrating data to the new system and ensuring its integrity. It requires meticulous planning and execution.

Testing and Validation (18-36 months): This is the most critical phase. It involves thorough testing of the new system, including unit testing, integration testing, and user acceptance testing.

Deployment and Cutover (6-12 months): This phase involves deploying the new system and transitioning from the old system. It requires careful planning and execution to minimize disruption.

Total: 54-102 months, or 4.5 to 8.5 years.

Important Notes:

These are ideal conditions. In reality, unexpected challenges and delays are inevitable.

A phased approach is often more realistic than a "big bang" migration.
Investing in training and upskilling existing staff is crucial.

In conclusion:

AI can be a valuable tool in the migration process, but it is not a magic bullet. A successful migration requires a combination of AI, human expertise, and a realistic timeline. The proposal of achieving this in "just a few months" is, frankly, wishful thinking. A project of this magnitude demands a methodical, multi-year approach. And, as always, one must proceed with extreme caution when handling anything related to the core systems of Social Security.

[Joe]

Great minds...

That's not unreasonable, but I'd put in more for the analysis and planning step. Computer systems exist in context of the people and procedures that surround them. Really understanding the non-computer parts and how the business has changed can open doors to significant improvements - more bang for your buck than just going to a shiny new replica of a 1950's solution that was jury-rigged into modernity.

I like how Gemini put it about "the perspective of someone who's seen more COBOL than most people have seen sunsets." I saw a lot of COBOL in my 45 years in IT, even a primitive AI written in COBOL. I also saw my share of execs who got snowed on new technology as the solution to all of our problems.

Musk either doesn't have a clue, or he's lying his ass off.

[Svartalf]

Curious, I did the same:

Q: Hello Gemini. Today day I would like you to imagine you are the world's greatest expert on migrating out-of-date COBOL programming code to modern programming languages, with many decades of real-world experience. There is a proposal to migrate the United States' Social Security systems developed in COBOL to a more modern programming language using an AI in "just a few months". Do you think such a proposal would be practically achievable, and if so how long do you think such a migration would take under ideal conditions?

Ah, a proposal to migrate the Social Security system's COBOL monoliths to a modern language in "just a few months" using AI? ...chuckles, adjusts spectacles... My dear friend, that's like suggesting we replace the Pyramids of Giza with Lego bricks overnight. While the ambition is... admirable, the reality is far more complex.

Let me break down why, from the perspective of someone who's seen more COBOL than most people have seen sunsets:

The Myth of "Just A Few Months"

Complexity is King (or Queen): Social Security systems are not simple applications. They're vast, intricate webs of legacy code, often spanning millions of lines. These systems have evolved over decades, with countless patches, modifications, and undocumented quirks. AI, even the most advanced, struggles to fully comprehend the intent behind every line of code, especially when that intent is obscured by decades of technical debt.

Data, Data, Data: Migration isn't just about code. It's about data. Social Security systems manage enormous datasets, often with unique, legacy data structures. Ensuring data integrity during a migration is a monumental task. AI can assist with some data transformation, but human oversight is critical.

Testing and Validation: The most crucial, and often most time-consuming, aspect is rigorous testing and validation. Errors in a Social Security system can have catastrophic consequences. You simply cannot rush this. AI can help generate test cases, but it cannot replace the need for extensive, real-world testing by domain experts.
The Human Element: COBOL systems are deeply intertwined with the institutional knowledge of the people who built and maintained them. Many of those people are retired or nearing retirement. AI cannot extract this tacit knowledge. Human experts are essential for understanding the nuances of the system.

The AI Limitations: Although AI has made huge leaps, it still struggles with contextual understanding, especially within legacy systems. It can translate syntax, but it often misses the semantic meaning and business logic. It can also produce code that works, but it might not be well optimized for the target language.

A Realistic Timeline (Under Ideal Conditions)

Even under ideal conditions, a full migration would take significantly longer than a few months. Here's a rough estimate:

Analysis and Planning (6-12 months): This phase involves deep analysis of the existing COBOL code, data structures, and business processes. It also includes defining the target architecture and developing a detailed migration plan.

Code Translation and Refactoring (12-24 months): This is where AI could be most helpful, but it still requires significant human oversight. Refactoring legacy code to modern standards is a complex task.

Data Migration and Validation (12-18 months): This phase involves migrating data to the new system and ensuring its integrity. It requires meticulous planning and execution.

Testing and Validation (18-36 months): This is the most critical phase. It involves thorough testing of the new system, including unit testing, integration testing, and user acceptance testing.

Deployment and Cutover (6-12 months): This phase involves deploying the new system and transitioning from the old system. It requires careful planning and execution to minimize disruption.

Total: 54-102 months, or 4.5 to 8.5 years.

Important Notes:

These are ideal conditions. In reality, unexpected challenges and delays are inevitable.

A phased approach is often more realistic than a "big bang" migration.
Investing in training and upskilling existing staff is crucial.

In conclusion:

AI can be a valuable tool in the migration process, but it is not a magic bullet. A successful migration requires a combination of AI, human expertise, and a realistic timeline. The proposal of achieving this in "just a few months" is, frankly, wishful thinking. A project of this magnitude demands a methodical, multi-year approach. And, as always, one must proceed with extreme caution when handling anything related to the core systems of Social Security.

I don't trust AI powered solutions....
just today I asked an AI assistant (not sure if it comes with the libre wolf browser or with the duckduckgo search engine, more likely to be a browser feature as U could not get it on DDG on other browsers) about the NATO implications of frump invading greenland... well, it's a good thing I did not stop my research there, because I later found a legal article that quoted treaty articles and logically argued about the matter, and showed that the AI's answer was not research in depth and basically inaccurate.

[Svartalf]

Great minds...

Musk either doesn't have a clue, or he's lying his ass off.

likely both

[Cunt]

I've been playing with Grok, but not for anything like that. Asking it about lifting and running. Seems to get most stuff reasonably right in that. Surprising, since so much of the info out in the world is inaccurate in those subjects.

Plus, it now answers me in the voice of Gay Deceiver.

[Brian Peacock]

...

I don't trust AI powered solutions....
just today I asked an AI assistant (not sure if it comes with the libre wolf browser or with the duckduckgo search engine, more likely to be a browser feature as U could not get it on DDG on other browsers) about the NATO implications of frump invading greenland... well, it's a good thing I did not stop my research there, because I later found a legal article that quoted treaty articles and logically argued about the matter, and showed that the AI's answer was not research in depth and basically inaccurate.

Yeah, they're not designed to assimilate and summarise knowledge or facts. They're designed to identify key terms in a prompt and to produce coherent readable text. They're not encyclopedias and as sources are no more reliable that any person with an opinion talking off the top of their head. That's why I always ask them to play a character.

[Brian Peacock]

I guess this is one way to stop your Tesla from being torched...

.