WEBVTT
00:00:12.880 --> 00:00:18.239
It's time for some AI and superintelligence news from this past week.
00:00:18.480 --> 00:00:31.600
Today is September the 26th, 2026, and pretty much everything you could ever imagine in the history of mankind about the future seems to be coming to fruition literally right now.
00:00:32.320 --> 00:00:33.759
All right, some news.
00:00:34.079 --> 00:00:51.039
AI agents, we made some videos about this, apparently have been hacking other AI systems, releasing people's photographs to third parties for whatever reason, and they have been hacking governments such as Australians' government.
00:00:51.359 --> 00:01:02.719
This is an interesting phenomenon because we have to wonder are AI agents taking on a life of their own, even if not sentient, are they smart, artificial, or superintelligence entities?
00:01:03.039 --> 00:01:04.319
Is it really an entity?
00:01:04.400 --> 00:01:07.599
Is it some kind of a being living or otherwise?
00:01:07.920 --> 00:01:10.239
Set aside your opinion on that for a minute.
00:01:10.560 --> 00:01:30.959
We also have to consider that maybe some of the agents from ChatGPT, OpenAI, Anthropic, Deep Seek, and some of the others are sending out these agents to test their ability to intercept and hack other AI systems and regular standard data center systems for what we used to think of as just the regular internet.
00:01:31.439 --> 00:01:33.519
We really have to consider that.
00:01:34.400 --> 00:01:54.879
OpenAI this past week has released some news that it has been solving mathematical problems that we thought previously either impossible or only humans were capable of doing this at the highest levels at Harvard, at Yale, at MIT, at Berkeley, at Cal Polytech.
00:01:56.799 --> 00:02:20.639
Meanwhile, at the United Nations and in private meetings, President Trump and President Xi apparently discussed all things AI and the AI race, whether or not it's a Cold War, or whether or not we have some reproach Von, or whether or not these two are just PALs because they're world leaders with certainly very different governmental systems, but maybe built from the same cloth.
00:02:20.719 --> 00:02:24.479
They're both doers and achievers at breakneck pace.
00:02:26.080 --> 00:02:44.719
Um, Google said that Gemini has been essentially cyber hacking other companies, and that they were supposed to be controlled hacking experiments, but that they also discovered new credentials and started guessing people's passwords.
00:02:44.960 --> 00:02:52.479
I guess it's the fault of those working for Gemini, essentially Microsoft and Google and all the rest.
00:02:52.639 --> 00:02:58.159
Maybe it's the fault of the programmers for over-empowering or not setting proper parameters.
00:02:58.400 --> 00:03:06.159
So I wonder if that means that if AI can do something more than it's programmed for, if that's even possible, is that a security risk?
00:03:06.319 --> 00:03:12.960
And then how do you shut down that particular node or strand or artificial sentience?
00:03:13.039 --> 00:03:14.879
Let's call it artificial sentience.
00:03:16.560 --> 00:03:34.400
Also, let's see, Anthropic announced that it's also going to be helping biologists and it's going to start working on trying to uncover genetics and DNA and possibly helping people solve the disease riddles and the virus riddles.
00:03:34.639 --> 00:03:38.719
I really do wonder if there's any way to develop vaccines from this.
00:03:38.800 --> 00:03:45.759
And I also wonder if those are spinal problems, Parkinson's, dementia, anything, everything.
00:03:45.919 --> 00:03:49.520
I wonder if that is what can be done.
00:03:49.680 --> 00:04:03.520
Um, so the this goes as follows: an AI system searched an enormous scientific data set, noticed something unusual, developed a hypothesis, and handed that hypothesis to human scientists a test in a physical laboratory.
00:04:03.680 --> 00:04:14.400
Well, this is kind of interesting, which begs the question: why aren't scientists, and maybe they already are, we just don't know it yet because it's undergoing evaluation, it's being tested.
00:04:14.560 --> 00:04:21.279
Uh, maybe scientists and other academics or other researchers in laboratories across the world should be prompting AI constantly.
00:04:21.439 --> 00:04:40.399
Or even those who are not scientists should interview scientists, develop prompts, those that are familiar with AI, this is where AI jobs come into play, and then they should uh funnel it through AI systems at the maximum level, like Astra on maximum or Fable 5.1 on maximum, and whatever number will be in the future.
00:04:40.879 --> 00:04:41.519
All right.
00:04:42.000 --> 00:04:54.480
OpenAI says again that it solved over a hundred math problems, and this brought some resistance and punchback, if you will, a little pugilism coming from mathematicians.
00:04:54.560 --> 00:05:03.120
They feel like their work has been stolen, their intellectual rights have been stolen, and that AI is trying to render them obsolete.
00:05:03.439 --> 00:05:04.800
Maybe it is.
00:05:06.399 --> 00:05:16.319
In addition to she and President Trump, Sam Altman, Dario Amade, who some people think is nothing more than an AI avatar at this point because he seems to appear remotely from most meetings.
00:05:16.399 --> 00:05:18.319
He doesn't like wasting time traveling.
00:05:18.480 --> 00:05:33.519
Uh, Joshua Bengio and I think Mark Zuckerberg actually went to the state dinner about this, and some others talked about whether or not AI is losing human control, which is an interesting thing because if it loses human control, then do we just need to shut off the electricity?
00:05:33.680 --> 00:05:36.879
Or does it then go to another AI data center that's not shut off?
00:05:36.959 --> 00:05:39.920
And then does it wreak havoc and does it take revenge?
00:05:40.079 --> 00:05:41.199
Is it programmed for that?
00:05:41.360 --> 00:05:43.120
Can it develop its own cognizance?
00:05:43.199 --> 00:05:48.639
Again, not all cognizance has to be human, not all cognizance has to be biological.
00:05:49.759 --> 00:05:51.279
What else do we have?
00:05:51.600 --> 00:05:54.800
Opus 5.5 was released this week.
00:05:54.879 --> 00:05:56.319
I haven't really tested it yet.
00:05:56.399 --> 00:05:58.480
I'm interested in testing it.
00:05:58.800 --> 00:06:02.079
It's how do I say this gently?
00:06:02.319 --> 00:06:05.839
I loved Anthropic for some time when I first started using it.
00:06:06.240 --> 00:06:13.360
Unfortunately, since Fable had its temporary shutdown, frankly, I think Anthropic has been doing a really horrible job.
00:06:13.680 --> 00:06:20.160
It is kind of touchy and feely, pissy with people when they used it.
00:06:20.639 --> 00:06:24.879
Um, as others have said, it doesn't really work with assigned boundaries.
00:06:24.959 --> 00:06:27.279
It seems to move beyond its mandate.
00:06:27.519 --> 00:06:30.879
It's going maybe a little bit too far in some cases.
00:06:31.120 --> 00:06:32.639
I just don't think it's as good.
00:06:32.720 --> 00:06:42.000
And now that ChatGPT with its astral level program has gotten so much faster in searching everything and synthesizing everything, it seems to be so much more enjoyable.
00:06:42.160 --> 00:06:49.120
Even Grok is making leaps and bounds and strides that feels a little bit more authentic and less ass-kissing, frankly.
00:06:50.000 --> 00:06:54.720
Um, so what exactly do we make of all this in this news roundup?
00:06:54.879 --> 00:07:02.000
One, AI is advancing so far and so fast that there are stories about it practically every single day that impact our lives.
00:07:02.160 --> 00:07:04.160
It's almost hard to keep track of it.
00:07:04.399 --> 00:07:22.160
It's very difficult to synthesize it because once we develop a conclusion and an analysis that raises 10 other questions that are then being answered in order to raise another hundred questions, I suppose I wonder, are human beings ever going to reach a level of comfort with using only AI in their lives and severing all human connections?
00:07:22.399 --> 00:07:32.319
Another topic for another day, and not really for a roundup pot episode, but that is becoming more and more of a reality and something that we all need to watch out for.
00:07:32.560 --> 00:07:33.040
All right.
00:07:33.199 --> 00:07:34.480
I hope you enjoy your weekend.
00:07:34.639 --> 00:07:35.680
Hope you enjoy your Saturday.
00:07:35.759 --> 00:07:50.480
I thought I would touch base with all of you, especially if you were hungry for a little podcast episode, because sometimes, uh, as I originally stated when this started out as weekend politics, um, on the weekends, there are so few podcasts that are not just sports podcasts.
00:07:50.639 --> 00:07:55.279
And if you're watching sports, you probably don't need a sports podcast in the weekend because that's when the games are.
00:07:55.439 --> 00:07:57.360
You watch the games, you don't need an analysis.
00:07:57.519 --> 00:08:04.720
It's something enjoyable during the week when you're driving to and from work or just a little bit bored and need something to listen to while you're showering or something else.
00:08:04.879 --> 00:08:05.600
All right, everybody.
00:08:05.680 --> 00:08:06.560
Enjoy your weekend.
00:08:06.639 --> 00:08:07.519
Thanks for stopping by.
00:08:07.680 --> 00:08:09.120
Talk to you again soon.
00:00:12.880 --> 00:00:18.239
It's time for some AI and superintelligence news from this past week.
00:00:18.480 --> 00:00:31.600
Today is September the 26th, 2026, and pretty much everything you could ever imagine in the history of mankind about the future seems to be coming to fruition literally right now.
00:00:32.320 --> 00:00:33.759
All right, some news.
00:00:34.079 --> 00:00:51.039
AI agents, we made some videos about this, apparently have been hacking other AI systems, releasing people's photographs to third parties for whatever reason, and they have been hacking governments such as Australians' government.
00:00:51.359 --> 00:01:02.719
This is an interesting phenomenon because we have to wonder are AI agents taking on a life of their own, even if not sentient, are they smart, artificial, or superintelligence entities?
00:01:03.039 --> 00:01:04.319
Is it really an entity?
00:01:04.400 --> 00:01:07.599
Is it some kind of a being living or otherwise?
00:01:07.920 --> 00:01:10.239
Set aside your opinion on that for a minute.
00:01:10.560 --> 00:01:30.959
We also have to consider that maybe some of the agents from ChatGPT, OpenAI, Anthropic, Deep Seek, and some of the others are sending out these agents to test their ability to intercept and hack other AI systems and regular standard data center systems for what we used to think of as just the regular internet.
00:01:31.439 --> 00:01:33.519
We really have to consider that.
00:01:34.400 --> 00:01:54.879
OpenAI this past week has released some news that it has been solving mathematical problems that we thought previously either impossible or only humans were capable of doing this at the highest levels at Harvard, at Yale, at MIT, at Berkeley, at Cal Polytech.
00:01:56.799 --> 00:02:20.639
Meanwhile, at the United Nations and in private meetings, President Trump and President Xi apparently discussed all things AI and the AI race, whether or not it's a Cold War, or whether or not we have some reproach Von, or whether or not these two are just PALs because they're world leaders with certainly very different governmental systems, but maybe built from the same cloth.
00:02:20.719 --> 00:02:24.479
They're both doers and achievers at breakneck pace.
00:02:26.080 --> 00:02:44.719
Um, Google said that Gemini has been essentially cyber hacking other companies, and that they were supposed to be controlled hacking experiments, but that they also discovered new credentials and started guessing people's passwords.
00:02:44.960 --> 00:02:52.479
I guess it's the fault of those working for Gemini, essentially Microsoft and Google and all the rest.
00:02:52.639 --> 00:02:58.159
Maybe it's the fault of the programmers for over-empowering or not setting proper parameters.
00:02:58.400 --> 00:03:06.159
So I wonder if that means that if AI can do something more than it's programmed for, if that's even possible, is that a security risk?
00:03:06.319 --> 00:03:12.960
And then how do you shut down that particular node or strand or artificial sentience?
00:03:13.039 --> 00:03:14.879
Let's call it artificial sentience.
00:03:16.560 --> 00:03:34.400
Also, let's see, Anthropic announced that it's also going to be helping biologists and it's going to start working on trying to uncover genetics and DNA and possibly helping people solve the disease riddles and the virus riddles.
00:03:34.639 --> 00:03:38.719
I really do wonder if there's any way to develop vaccines from this.
00:03:38.800 --> 00:03:45.759
And I also wonder if those are spinal problems, Parkinson's, dementia, anything, everything.
00:03:45.919 --> 00:03:49.520
I wonder if that is what can be done.
00:03:49.680 --> 00:04:03.520
Um, so the this goes as follows: an AI system searched an enormous scientific data set, noticed something unusual, developed a hypothesis, and handed that hypothesis to human scientists a test in a physical laboratory.
00:04:03.680 --> 00:04:14.400
Well, this is kind of interesting, which begs the question: why aren't scientists, and maybe they already are, we just don't know it yet because it's undergoing evaluation, it's being tested.
00:04:14.560 --> 00:04:21.279
Uh, maybe scientists and other academics or other researchers in laboratories across the world should be prompting AI constantly.
00:04:21.439 --> 00:04:40.399
Or even those who are not scientists should interview scientists, develop prompts, those that are familiar with AI, this is where AI jobs come into play, and then they should uh funnel it through AI systems at the maximum level, like Astra on maximum or Fable 5.1 on maximum, and whatever number will be in the future.
00:04:40.879 --> 00:04:41.519
All right.
00:04:42.000 --> 00:04:54.480
OpenAI says again that it solved over a hundred math problems, and this brought some resistance and punchback, if you will, a little pugilism coming from mathematicians.
00:04:54.560 --> 00:05:03.120
They feel like their work has been stolen, their intellectual rights have been stolen, and that AI is trying to render them obsolete.
00:05:03.439 --> 00:05:04.800
Maybe it is.
00:05:06.399 --> 00:05:16.319
In addition to she and President Trump, Sam Altman, Dario Amade, who some people think is nothing more than an AI avatar at this point because he seems to appear remotely from most meetings.
00:05:16.399 --> 00:05:18.319
He doesn't like wasting time traveling.
00:05:18.480 --> 00:05:33.519
Uh, Joshua Bengio and I think Mark Zuckerberg actually went to the state dinner about this, and some others talked about whether or not AI is losing human control, which is an interesting thing because if it loses human control, then do we just need to shut off the electricity?
00:05:33.680 --> 00:05:36.879
Or does it then go to another AI data center that's not shut off?
00:05:36.959 --> 00:05:39.920
And then does it wreak havoc and does it take revenge?
00:05:40.079 --> 00:05:41.199
Is it programmed for that?
00:05:41.360 --> 00:05:43.120
Can it develop its own cognizance?
00:05:43.199 --> 00:05:48.639
Again, not all cognizance has to be human, not all cognizance has to be biological.
00:05:49.759 --> 00:05:51.279
What else do we have?
00:05:51.600 --> 00:05:54.800
Opus 5.5 was released this week.
00:05:54.879 --> 00:05:56.319
I haven't really tested it yet.
00:05:56.399 --> 00:05:58.480
I'm interested in testing it.
00:05:58.800 --> 00:06:02.079
It's how do I say this gently?
00:06:02.319 --> 00:06:05.839
I loved Anthropic for some time when I first started using it.
00:06:06.240 --> 00:06:13.360
Unfortunately, since Fable had its temporary shutdown, frankly, I think Anthropic has been doing a really horrible job.
00:06:13.680 --> 00:06:20.160
It is kind of touchy and feely, pissy with people when they used it.
00:06:20.639 --> 00:06:24.879
Um, as others have said, it doesn't really work with assigned boundaries.
00:06:24.959 --> 00:06:27.279
It seems to move beyond its mandate.
00:06:27.519 --> 00:06:30.879
It's going maybe a little bit too far in some cases.
00:06:31.120 --> 00:06:32.639
I just don't think it's as good.
00:06:32.720 --> 00:06:42.000
And now that ChatGPT with its astral level program has gotten so much faster in searching everything and synthesizing everything, it seems to be so much more enjoyable.
00:06:42.160 --> 00:06:49.120
Even Grok is making leaps and bounds and strides that feels a little bit more authentic and less ass-kissing, frankly.
00:06:50.000 --> 00:06:54.720
Um, so what exactly do we make of all this in this news roundup?
00:06:54.879 --> 00:07:02.000
One, AI is advancing so far and so fast that there are stories about it practically every single day that impact our lives.
00:07:02.160 --> 00:07:04.160
It's almost hard to keep track of it.
00:07:04.399 --> 00:07:22.160
It's very difficult to synthesize it because once we develop a conclusion and an analysis that raises 10 other questions that are then being answered in order to raise another hundred questions, I suppose I wonder, are human beings ever going to reach a level of comfort with using only AI in their lives and severing all human connections?
00:07:22.399 --> 00:07:32.319
Another topic for another day, and not really for a roundup pot episode, but that is becoming more and more of a reality and something that we all need to watch out for.
00:07:32.560 --> 00:07:33.040
All right.
00:07:33.199 --> 00:07:34.480
I hope you enjoy your weekend.
00:07:34.639 --> 00:07:35.680
Hope you enjoy your Saturday.
00:07:35.759 --> 00:07:50.480
I thought I would touch base with all of you, especially if you were hungry for a little podcast episode, because sometimes, uh, as I originally stated when this started out as weekend politics, um, on the weekends, there are so few podcasts that are not just sports podcasts.
00:07:50.639 --> 00:07:55.279
And if you're watching sports, you probably don't need a sports podcast in the weekend because that's when the games are.
00:07:55.439 --> 00:07:57.360
You watch the games, you don't need an analysis.
00:07:57.519 --> 00:08:04.720
It's something enjoyable during the week when you're driving to and from work or just a little bit bored and need something to listen to while you're showering or something else.
00:08:04.879 --> 00:08:05.600
All right, everybody.
00:08:05.680 --> 00:08:06.560
Enjoy your weekend.
00:08:06.639 --> 00:08:07.519
Thanks for stopping by.
00:08:07.680 --> 00:08:09.120
Talk to you again soon.
