re: Anyone else read the “AI 2027” Scenario?

quote:

---

The computer is not "reasoning" the same way we do, it is just better at parsing large data sets better than what we are accustomed to

---

I use ChatGPT as a personal assistant and it is amazing. It took me a few months to code and trouble shoot it the way I want. It is definitely not “intelligent.”

I think between people ascribing humanity to it and just ignorance they are too quick to claim AI is actually intelligent. It’s not.

That being said it is absolutely an economic and social disruption.

AI can and will replace a ton of jobs especially menial jobs like data entry and analytics and marketing, etc.

You’ll still ultimately need a human to review the data but even right now in my industry what used to take a person days in data review now takes hours and I can see that getting down to minutes in the near future.

The much much bigger problem is the evil class of power hungry lunatics trying to use learning models for their own goals like authoritarian fascism and social revolution and to that end “AIs” ability to algorithmically manipulate society is absolutely terrifying.

quote:

---

We’re going to see a plateau in advancement sometime in the near future most likely

---

They’ll use AI to perfect nuclear power plants and then build them next to the data centers.

The problem is the absolutely absurd amounts of power and water these data centers demand.

quote:

---

dont think anyone really understands that LLMs are not really true AI in the way normal people use term (science fiction movies). The computer is not "reasoning" the same way we do, it is just better at parsing large data sets better than what we are accustomed to.

---

Absolutely. It’s still a mathematical model at this point. But we have people thinking they are dating the LLM they talk to, so we are already seeing peak hysteria.

I don’t understand the experts theorizing a relatively instant breakthrough from the AI we currently have to AGI or theorized ASI. All the extremes jump out to us rapidly developing AI with no hiccups and a smooth exponential growth curve, including the leaps and bounds to go from a mathematical model to a thinking and reasoning ASI. There will be a large plateau in development where we hit a wall, potentially based on power like you suggested, or based on hitting a temporary human creative wall

quote:

---

The disruption in the coding world may be accurate.

---

IF "AI" continues to develop as predicted, not this AGI sentient bullshite, but as simply a technological advancement, all white collar jobs are on the table in phase 1.

Phase 1:
The soft stage

Lawyers
Insurance Underwriters
IT Support
Coders
Project Managers
Business Analysts
Data Analysts
Financial Analysts
Tax Consultants
Accountants
Alot of Engineers
Architects
Systems Analysts
Data Scientists
Most Cybersecurity Analysts
Alot of mid level diagnostic medical positions
etc, etc

And ironically, the ones easiest to replace - all the suits in the c-suites

Phase 2:
The medical device, digital currency, UBI, simple robotics stage: once the AI companies consolidate billions/trillions based on their absurd subscription models from replacing the above and crashing the economy lowering the costs of everything while healthcare dwindles, the next wave of advances will revolve around giving those who have lost the jobs above a new place to work in robotics. Assuming governments aren't bought and paid off by these companies, they will try to gain control of the AI industry as a "public service" to maintain order. Either way, cheap robotic labor/tech will be developed that will replace:

Almost all mid-level medical providers
Almost all diagnostic medical providers
Almost all claims clearing-houses
Almost all insurance processing
Almost all billing and tax processing across the board
Police forces will shrink on the patrol side

At this point, some idiots will have successfully pushed for a digital currency because that was always the goal.

We will then be pushed towards UBI with social credit scores used to maintain order

Phase 3:
The advanced robotics, international cyber-warfare stage

Assuming society doesn't completely collapse by this point, this is the advanced robotics wave

Most Wielders
Most Surgeons
etc etc

This will come with advanced warfare we won't be prepared for.

The limiting factors will be metals, energy, the knowledge bootstrap problem, and the human capacity to push forward with this bullshite

I mean if Hollywood writers,dock workers , autoworker unions etc... all want to unplug A.I. to save jobs then why not everyone else?

So... Skynet

LLMs are just highly sophistical language prediction systems. They don't understand anything. Think "google search autocomplete" on steroids. LLMs have to incorporate with multiple agents and other more complex reasoning systems (like code interpreters, calculators, etc) in hybrid architectures to advance further, as of today.

Ultimately, the limit of any "AI" system is going to be the human knowledge that feeds it, making it difficult for AI surpass human knowledge.There would be no verification process other than a self-referential, possibly incoherent, one for the AI system.

quote:

---

Does this mean that the less we put on the internet, the harder it will be for AI to advance?

---

Pretty much

quote:

---

Can Asimov’s Three Laws of Robotics save us?

---

They don't apply

Current LLM systems can’t meaningfully understand what harm is, they don’t actually obey in any intentional sense because they just generate responses based on statistical patterns, and have no concept of self-preservation or existence because they’re stateless text generators

quote:

---

• Job Disruption: By 2027, AI automates coding and research, reducing demand for programmers and researchers. Industries like customer service and data analysis face upheaval.

---

Too bad for all those "learn to code" a-holes