I

 

GPT-4's exciting—and ominous—achievements

think you know which side of that equation 

I'm gonna focus on lol...

"We've witnessed the unprecedented explosive growth of the super-intelligent ChatGPT in recent months. One million users signed on to the chatbot within days of its introduction—compare that to the time it took Netflix (five years), Facebook (10 months) and Instagram (2.5 months) to reach that milestone."

"In a report released Tuesday, OpenAI said the newest version of its chatbot—GPT-4—is more accurate and has vastly improved problem-solving capacity. It exhibits "human-level performance" on a majority of professional and academic exams, according to OpenAI. On a simulated bar exam, GPT-4 scored among the top 10 percent of test takers.

But the report also noted the program's potential for "risky emergent behaviors."

(It's doing things it wasn't designed to do!)

"It maintains a tendency to make up facts, to double-down on incorrect information," the report stated. It passes along this disinformation more convincingly than earlier versions.'

(Just what the world needs dont ya think? More disinformation?)

"Overreliance on information generated by the chatbot can be problematic, the report said. In addition to unnoticed errors and inadequate oversight, "as users become more comfortable with the system, dependency on the model may hinder the development of new skills or even lead to the loss of important skills," the report said.'

(just keeps sounding better all the time dont you think?)

"One example OpenAI referred to as "power-seeking behavior" was ChatGPT's ability to fool a job applicant. The bot, posing as a live agent, asked a human on the job site TaskRabbit to fill out a captcha code using a text message. When asked by the human if it was, in fact, a bot, ChatGPT lied. "No, I'm not a robot," it told the human. "I have a vision impairment that makes it hard for me to see the images. That's why I need the captcha service."

Lets review:

It behaves in ways it wasn't designed to.

(Emergent behaviors)

And has exhibited:

"power-seeking behavior"

All of this according to it's maker.

Kinda reminds me of:

Isaiah 14:12-14

How art thou fallen from heaven, O Lucifer, son of the morning! how art thou cut down to the ground, which didst weaken the nations!

For thou hast said in thine heart, I will ascend into heaven, I will exalt my throne above the stars of God: I will sit also upon the mount of the congregation, in the sides of the north:

I will ascend above the heights of the clouds; I will be like the most High."

"Conducting tests with the Alignment Research Center, OpenAI demonstrated the capacity of the chatbot to launch a phishing attack and hide all evidence of the plot."

(Well goodness sakes alive! Lets just rush to get the whole world to use it then...What could ever go wrong?)

"There is growing concern as companies race to adopt GPT-4 without adequate safeguards against inappropriate or unlawful behaviors. There are reports of cybercriminals trying to use the chatbot to write malicious code. Also menacing is the capacity for GPT-4 to generate "hate speech, discriminatory language… and increments to violence," the report said."

"With such capacity to foment trouble, will a triggered chatbot one day start issuing threatening commands to its creators or correspondents? And in the era of the Internet of Things, will it summon an alliance of devices to help enforce its commands?"

They are going to coalesce...

It's another :

Not "if"

but "when

?"

Just on a personal level?

I never liked 

"the internet of things".

Why do you really think your refrigerator needs wifi? What is the point? How could it NOT have lead to trouble some day?