I had a single "C" class in engineering school and did a little programming on a microcontroller for robot competition, but that's it. A couple of weeks ago I asked Grok to do a microcontroller program and specified it to be in Python, along with an interface so I could run the device from my Mac.
It took less than five seconds for the program to arrive in my inbox. Haven't run it yet, but a relative - a mainframe programmer for an insurance company - needed a module, did the very same thing and said it ran perfectly. I expect no less.
Per Oscar Wilde a cynic knows the cost of everything, but the value of nothing…one of my favorites.
I few years ago I wrote a book section that pointed out the irony of an analogy few noticed i.e. the Mars Rover ‘Spirit’ is not dead in that every bit of data collected was stored in a memory bank in the Jet Propulsion Laboratory in Pasadena, California…it has a serial number and an identical rover could be constructed and the data downloaded e.g. DNA and memory stored in an early “cloud” in effect resurrecting the dead.
I also included a few examples of savants across history that had inexplicable talent and skills that have no explanation save for what scientists have come to realize…individual brain cells that are remarkably similar to quantum computers with the ability to share information in a very real “cloud”.
It has always struck me that those who have made fortunes sequencing binary code fail to realize they are walking on the lawn of the Creator of a binary existence.
Glenn: While your article is totally spot on (regardless of some "boob" comments!) I wonder about AI being used to deliberately alter well known persons and have them "supposedly say" things that we wouldn't expect them to say of have such ideas. Also, how about re-creating movies, etc. with different people as actors? Just wondering.
It's a real problem already. The Pope has a long AI sermon, Jordan Peterson has many talks on YuTube. My son is an AI engineer. He said if it isn't live, with those people, it's AI cashing in on views by faking popluar people. Even Barron Trump is shown singing fake songs.
Yes, You Tube is awash with fake discourses: Are there no regulations or ethical guidelines determining what can appear there? I seem to remember that You Tube has censored certain kinds of content in recent history, but it’s the Wild West currently.
A fair point, but isn’t there a difference between censoring points of view and disallowing what are obviously fake, deceptive portrayals of public figures when there is no redeeming satirical justification for such distortions of their words and actions?
Yes, let the consumer of this junk beware, but why do we even have an FCC then?
I think we should NOT have an FCC. Politicians cannot resist the urge to politicize it, and it's not really needed in an age of infinite bandwidth.
As for fake depictions of Elon Musk saying fake things, (1) the fact that you recognize that they're fake suggests they aren't really harmful in a fraudulent sense, and (2) Elon has ample remedies in civil law.
Yes. Me, too. However, taking the libertarian (small "l") point of view, the gub'ment's primary function is to protect the weak from the strong and to do that, men make laws. To my eye this is fraud - using someone famous' likeness and voice without compensation - fraud not only committed against that person but every person who views the video with the expectation of benefitting from their wisdom.
Indeed, it's fraud. But the primary remedy for fraud is a civil action. Elon has that civil action available to him (meaning his army of lawyers) if he feels like pursuing it.
And I suppose the Pope has lawyers as well, but how do you sue an AI algorithm coming from God knows where? And aren’t the courts tied up with enough nonsense as it is?
I'm seeing a lot of them with Elon. They haven't mastered the head motion yet, so if you look for that, you can spot them. It irritates me, though, that they are using his likeness and voice to hawk their own points of view and make money with YouTube's algorithm. Another of many disturbing Google actions.
I can. I completely stopped using ChatGPT. I had posed several questions asking it to look deeper into and explain various historical events, and it came back with the most sickening lectures on DIE (not DEI - DIE). So as a test, I asked it about several well-known historical quotes (Dr.Johnson's about women preachers being like dogs walking on their hind legs, for example), and it went off the deep end.
That day I subscribed to Grok, and only once has it asserted to me anything remotely - and I mean remotely - Woke; I asked it to permanently stop doing that, and so far it has.
Grok spewed some nonsense at me about men who think they are women and gender reveal parties. I corrected it and it totally caved to my non woke viewpoint.
In another session, discussing words with four or more o's, nonhomologous oophorotomy it totally miscounted all the words o's. Bizarre! I made it stop that mode by a program modification. Basic.Oophorocoloprostostmy was its final try, not quite a real word. Lower sigmoid colon removal including ovaries.
I get it. Haven't seen too many modern women who didn't have their ovaries up their ..... well, you know. Especially those in Congress and the Senate. Must be a requirement of DIE.
Thank you. I asked Grok (free version) to tell me about left wing vs. right wing violence. It told me that right wing violence was 5x more common than left wing. I made every effort I knew how to make to make the query objective and accurate. Seems to me GROK is also relying on left-leaning sources. Maybe these are all that's available? Note that I asked GROK to only use sources it considered reliable. You can read the entire conversation at : https://grok.com/share/c2hhcmQtMw%3D%3D_72991fd7-853d-452e-8477-b768361c720c Please do read it before replying, if indeed, you wish to reply.
It gives false answers about political talking points about violence, chat GTP told me that right wIng violence exceeds any left wing violence, and that is just not true
I had a conversation with an AI program recently regarding Trump and Russiagate. AI said it was not a hoax that the allegations were true. After multiple questions, it stated that Russiagate was likely 99% false and possibly there was some small trace of truth in there.
If AI is truly’thinking’, it’s not thinking clearly on many topics….
Still though, in my job advising top execs on Financial matters, I found that gathering and organizing data was onerously difficult. Generally I needed one or more employees and at least $100k in subscriptions to financial databases… AI can probably do that work for far less.
I'm hardly the first to notice that AI (not just Chat GPT) seems to be programmed to suck up to people. It will tell us what it thinks we want to hear. But it may be wrong about what we want to hear. I've found that instructing it to not compliment me and not agree with me unless it is 95% certain I'm right helps mitigate this tendency. Also, telling it to think long and hard before agreeing with me helps. And instructing it to only agree when it can cite at least 3 sources for its agreement.
I used to think that if enough people applied enough brainpower, including micro-slicing a human brain and meticulously tracing every axon, we would eventually understand it.
I have changed my mind. There is something fundamental that we don't remotely understand. Ours is a computing engine using about 25 watts; AI centres use about 200 million watts to imitate poorly the nuances of that 25W machine. We're missing something. A wasp has about a million neurons, yet it can fly, stay upright and navigate while dealing with wind and find its way home. It can find food, build nests, reproduce. I have some robotics experience; we can't touch that wasp's capabilities with billions of transistors.
Stargazing, I often think of those early-historical shepherds sleeping with their flocks and looking at the same stars, wondering what they were, where they were, and what made them shine. The finest hypothetical science-fiction writer of their day could not conceive of the distance, let alone the mechanism of their luminance. Not even remotely.
We are very likely at that same point. There is fundamental physics of which we have no clue; one current theory is that animal thinking has a quantum-computing component.
I'd love to live another thousand years. We are but babes in this game - I would like to see how it turns out.
AI is changing so fast, our son was here giving a zoom talk to Nvidia colleagues about their latest tool, entirely different than July's one. They are very excited and it just came out. They use tools to make tools. Exponentially.
I had a single "C" class in engineering school and did a little programming on a microcontroller for robot competition, but that's it. A couple of weeks ago I asked Grok to do a microcontroller program and specified it to be in Python, along with an interface so I could run the device from my Mac.
It took less than five seconds for the program to arrive in my inbox. Haven't run it yet, but a relative - a mainframe programmer for an insurance company - needed a module, did the very same thing and said it ran perfectly. I expect no less.
Pretty scary.
You are a boob. AI does not think
I'm curious. Is your comment a representative example of your personal "thinking?"
No , it is a reflection on your thinking - such as it is.
I'm relieved that your comment is not an example of your thinking.
Now that we've resolved our disagreement, feel free to take a moment to read the piece. You can afford it - it's free!
You're making me laugh! Don't stop!
I'm curious about another thing, however. Is "boob" your only epithet?
Per Oscar Wilde a cynic knows the cost of everything, but the value of nothing…one of my favorites.
I few years ago I wrote a book section that pointed out the irony of an analogy few noticed i.e. the Mars Rover ‘Spirit’ is not dead in that every bit of data collected was stored in a memory bank in the Jet Propulsion Laboratory in Pasadena, California…it has a serial number and an identical rover could be constructed and the data downloaded e.g. DNA and memory stored in an early “cloud” in effect resurrecting the dead.
I also included a few examples of savants across history that had inexplicable talent and skills that have no explanation save for what scientists have come to realize…individual brain cells that are remarkably similar to quantum computers with the ability to share information in a very real “cloud”.
It has always struck me that those who have made fortunes sequencing binary code fail to realize they are walking on the lawn of the Creator of a binary existence.
Glenn: While your article is totally spot on (regardless of some "boob" comments!) I wonder about AI being used to deliberately alter well known persons and have them "supposedly say" things that we wouldn't expect them to say of have such ideas. Also, how about re-creating movies, etc. with different people as actors? Just wondering.
It's a real problem already. The Pope has a long AI sermon, Jordan Peterson has many talks on YuTube. My son is an AI engineer. He said if it isn't live, with those people, it's AI cashing in on views by faking popluar people. Even Barron Trump is shown singing fake songs.
Yep. And of course, even if it purports to be live, that's no guarantee it's real.
Yes, You Tube is awash with fake discourses: Are there no regulations or ethical guidelines determining what can appear there? I seem to remember that You Tube has censored certain kinds of content in recent history, but it’s the Wild West currently.
Caveat emptor.
The last time we had "rules" about such things was during the Biden administration. I'll take the Wild West over the Socialist East.
A fair point, but isn’t there a difference between censoring points of view and disallowing what are obviously fake, deceptive portrayals of public figures when there is no redeeming satirical justification for such distortions of their words and actions?
Yes, let the consumer of this junk beware, but why do we even have an FCC then?
I think we should NOT have an FCC. Politicians cannot resist the urge to politicize it, and it's not really needed in an age of infinite bandwidth.
As for fake depictions of Elon Musk saying fake things, (1) the fact that you recognize that they're fake suggests they aren't really harmful in a fraudulent sense, and (2) Elon has ample remedies in civil law.
Yes. Me, too. However, taking the libertarian (small "l") point of view, the gub'ment's primary function is to protect the weak from the strong and to do that, men make laws. To my eye this is fraud - using someone famous' likeness and voice without compensation - fraud not only committed against that person but every person who views the video with the expectation of benefitting from their wisdom.
Indeed, it's fraud. But the primary remedy for fraud is a civil action. Elon has that civil action available to him (meaning his army of lawyers) if he feels like pursuing it.
And I suppose the Pope has lawyers as well, but how do you sue an AI algorithm coming from God knows where? And aren’t the courts tied up with enough nonsense as it is?
I'm seeing a lot of them with Elon. They haven't mastered the head motion yet, so if you look for that, you can spot them. It irritates me, though, that they are using his likeness and voice to hawk their own points of view and make money with YouTube's algorithm. Another of many disturbing Google actions.
I keep telling everyone that Google needs to change its motto from, "Don't be evil" to "Don't be Google." But I repeat myself.
Chat GTP is completely "Woke" and leftist
Can you give specifics that lead you to this conclusion?
I can. I completely stopped using ChatGPT. I had posed several questions asking it to look deeper into and explain various historical events, and it came back with the most sickening lectures on DIE (not DEI - DIE). So as a test, I asked it about several well-known historical quotes (Dr.Johnson's about women preachers being like dogs walking on their hind legs, for example), and it went off the deep end.
That day I subscribed to Grok, and only once has it asserted to me anything remotely - and I mean remotely - Woke; I asked it to permanently stop doing that, and so far it has.
Grok spewed some nonsense at me about men who think they are women and gender reveal parties. I corrected it and it totally caved to my non woke viewpoint.
In another session, discussing words with four or more o's, nonhomologous oophorotomy it totally miscounted all the words o's. Bizarre! I made it stop that mode by a program modification. Basic.Oophorocoloprostostmy was its final try, not quite a real word. Lower sigmoid colon removal including ovaries.
I get it. Haven't seen too many modern women who didn't have their ovaries up their ..... well, you know. Especially those in Congress and the Senate. Must be a requirement of DIE.
Claude Code. Look it up. It's the new mega AI that developers are excited about.
Thank you. I asked Grok (free version) to tell me about left wing vs. right wing violence. It told me that right wing violence was 5x more common than left wing. I made every effort I knew how to make to make the query objective and accurate. Seems to me GROK is also relying on left-leaning sources. Maybe these are all that's available? Note that I asked GROK to only use sources it considered reliable. You can read the entire conversation at : https://grok.com/share/c2hhcmQtMw%3D%3D_72991fd7-853d-452e-8477-b768361c720c Please do read it before replying, if indeed, you wish to reply.
It gives false answers about political talking points about violence, chat GTP told me that right wIng violence exceeds any left wing violence, and that is just not true
But, but, but, Biden said so!
I had a conversation with an AI program recently regarding Trump and Russiagate. AI said it was not a hoax that the allegations were true. After multiple questions, it stated that Russiagate was likely 99% false and possibly there was some small trace of truth in there.
If AI is truly’thinking’, it’s not thinking clearly on many topics….
Still though, in my job advising top execs on Financial matters, I found that gathering and organizing data was onerously difficult. Generally I needed one or more employees and at least $100k in subscriptions to financial databases… AI can probably do that work for far less.
I'm hardly the first to notice that AI (not just Chat GPT) seems to be programmed to suck up to people. It will tell us what it thinks we want to hear. But it may be wrong about what we want to hear. I've found that instructing it to not compliment me and not agree with me unless it is 95% certain I'm right helps mitigate this tendency. Also, telling it to think long and hard before agreeing with me helps. And instructing it to only agree when it can cite at least 3 sources for its agreement.
I used to think that if enough people applied enough brainpower, including micro-slicing a human brain and meticulously tracing every axon, we would eventually understand it.
I have changed my mind. There is something fundamental that we don't remotely understand. Ours is a computing engine using about 25 watts; AI centres use about 200 million watts to imitate poorly the nuances of that 25W machine. We're missing something. A wasp has about a million neurons, yet it can fly, stay upright and navigate while dealing with wind and find its way home. It can find food, build nests, reproduce. I have some robotics experience; we can't touch that wasp's capabilities with billions of transistors.
Stargazing, I often think of those early-historical shepherds sleeping with their flocks and looking at the same stars, wondering what they were, where they were, and what made them shine. The finest hypothetical science-fiction writer of their day could not conceive of the distance, let alone the mechanism of their luminance. Not even remotely.
We are very likely at that same point. There is fundamental physics of which we have no clue; one current theory is that animal thinking has a quantum-computing component.
I'd love to live another thousand years. We are but babes in this game - I would like to see how it turns out.
AI is changing so fast, our son was here giving a zoom talk to Nvidia colleagues about their latest tool, entirely different than July's one. They are very excited and it just came out. They use tools to make tools. Exponentially.