The man who predicted the 2008 crash, Anthropic’s co-founder, and a leading AI podcaster jump into a Google doc to debate the future of AI—and, possibly, our lives
AI revolution is a bunch of rich people trying to be puppet masters as they make every day citizens dumber while convincing them they are empowering them, while at the same time destabilizing the economy and destroying the environment. Data centers are poisoning people and the earth. All it’s doing is pushing us closer to a new Great Depression and dust bowl.
Substack is built on the foundation of human work, artistry and writing. Your whole platform wouldn’t exist if AI took over. Why would you platform something that is coming for the creative sector, while also killing the planet??
As a co inventor of the Internet and there were many of us starting in the 1960s; my contribution early 80s, was for a selfish reason to promote the Canadian film industry for which I was college trained yet underemployed.
I also knew it was a way to erase ignorance but now it basically spreads it, except for sites such as here and intelligent comments at youthtube.
So. What are you going to do? Internet and AI make a lot of people confused with overwhelming data who then select what they wish or know that they don't know what to use. Not everyone but most. So. What are you going to do?
Surprise, Surprise. More AI fear-porn. Honestly, AI gets nothing right, but uses eloquent speech and perfect grammar in its attempt to get you to believe the misinformation it just produced.
If you haven't figured it out by now, the predator-class wants you fearful, and AI is the next tool to get you scared. Don't play along. AI is retarded, and will always be a tool to regurgitate the crap it finds online.
If we refuse to read any more false prognostications of the end-of-human civilization due to AI, THEY WILL HAVE TO STOP WRITING ABOUT IT.
Exactly! AI’s real weakness is that it lacks imagination and true rationale. Those two fundamental elements are needed for it to be anything close to a human. Perhaps that’s exactly why AI is being developed the way it currently is. 🤔
This is where most people get AI wrong. It's not about subbing out YOUR imagination or rationale. We must always keep a human-in-the-loop; its essential for quality use. AI, as it is now, is a TOOL. A hammer can't hit a nail unless someone who knows how to swing it wields it. Same with AI. Sure, input a weak-ass prompt and it'll output slop; but input a well-thought out, highly detailed prompt and your output mirrors that.
Far too many people throw out the baby with the bath water because they don't understand what they're using.
Have you read about the recent existential dread from coders on how Claude code is taking away their flow activity of writing code and making them supervisors and manager. The last Claude model is much more than a tool it seems.
Did you read the article? It hardly constitutes "AI fear porn"; it's actually a well balanced, thoughtful and highly informed piece by some serious heavy hitters in finance, tech and AI.
Also, "retarded" is passé.
And, AI is an incredibly valuable tool, that gets a lot right, when used correctly. Just check out this prompt I created:
Daniel! You have assumed incorrectly. With 30 years of software development and systems support & analysis behind me, I have no fear of technology and a better understanding than most of what software can and cannot do. My irritation is the constant prognostication of what AI is going to do, from people who really don't know where it's going but are happy to weigh in with their authoritative-sounding guesses and stir up undue fear.
Your definition of "heavy hitters" is suspect. It's a designation that is historically used to amplify regime stooges from the dubious "expert-class" who are paid to create a usually false narrative designed to profit the connected.
I have no dog in the fight regarding these gentlemen, but I see the worry in those around me and how it is endlessly poked, and I am done with it. There seems to be something about new tech that makes a certain type of man feel the need to puff out his chest and go on the record as in-the-know. It's old already.
Oh, and honestly, who cares what YOU think is passé? When EVERY time I use AI it plies me with WRONG answers, that qualifies it for retardation status. And, I hardly think amplifying it as an "incredible value" because it will sometimes deliver something you can make use of, despite its habit of abject falsity, is an impressive flex. Every query offers the system another swipe at deceiving you.
Sorry for you that your guard is already fully down. But I retract nothing from my prior comment.
"With 30 years of software development and systems support & analysis behind me": Well, you sure hell don't sound like it in the comment I replied to earlier. Not to mention, you don't even sound like an adult with language like "AI is retarded." So if you want to be taken seriously with your experience -- act like it!
Is being an adult the minimum requirement for commenting here? Regardless my feelings are not hurt at your attempted insult. Psht.
You do realize tho that you are having an online tantrum about my assessments of AI and the men who are fan-boys of the sci-fi prognosticators. I could choose a different insulting term but I suspect you would complain about that one too.
At least acknowledged that you have internalized a need to feel savy about AI and therefore feel personal pain when it is insulted.
Karen! "Heavy-hitters" is baseball term later extended idiomatically to mean any very influential, important, or successful person or organization in business, politics, or other fields. It signifies someone with significant impact, power, or achievement, much like a baseball player's powerful hits drive runs for their team. So, not suspect; apt.
And, how exactly is a co-founder of Anthropic not in a place to "prognosticate on where AI is going"? Who is then?
What really is suspect is this: You have 30 years in software development yet you can't get a GenAI tool to provide you ANY value? Seriously? Or, are you just scared for you career yet refusing to adopt it and then projecting that self-made frustration out into an already fear filled world?
You don't need to retract anything, but to claim they were fear mongering in this piece is utterly hypocritical given your initial comment.
Wrong too bro! I am retired; not remotely threatened. When the first ten queries return falsity, there is no expertise that can be applied to rectify that. But it takes wisdom to discern that what one is looking at is erudite sounding Google-search results. And in case you are blissfully ignorant, Google search results are a combination of Intelligence-slop and paid advertisements masquerading as return data.
Spare me your definition of terms that don’t move the ball in the debate. Accolades are almost always propaganda, ladled liberally over stooges, regardless of the etymology of the term.
We live in different realities. You’ll wake up and figure it out sooner or later.
You are wildly misinformed -- like 99.99% of the people on here (and pretty much everywhere). Be it AI or anything else -- you never adjust in the light of new information (unlike AI): https://copilotandcaptain.com/
We do NOT have Artificial Intelligence. We have large language models that are basically like Google search on steroids. But we do not have Thinking Machines. "AI" is the worst misnomer of the 21st century.
I accidentally started a small coding group of people (most laid off) who want to build non-AI software. We started with a document distribution platform since kindle has made the pdf and epub format unsafe to distribute. Sorry about your job man. Fuck these guys.
I don't know why I was sent this substack post. I don't subscribe and this perspective is not one I support. I want no part of this please stop sending to me.
This AI craze is only “a revolution” because of the tech oligarchs and the Grifter in the Oval Office. AI is so far is a failed error prone model. It may help cheating college students on their papers and giving corporations availability to layoff majority of their white collar employees mistakenly confident that AI will replace humans. The new AI phone system at my pharmacy won’t allow me to request a refill like the previous analog DIY phone system. SIRI AI is annoyingly dunce that Apple has to use Google’s AI. Meanwhile, billions are being invested in these 24/7 data centers sucking up valuable fresh water and making residence pay for the subsidize electricity to them. Don’t reason why the stock market hit 49K is because of this AI bubble.
A lot of companies are quietly admitting that they were premature in laying people off in lieu of AI. They have found that AI doesn’t actually replace human productivity. We are going to see a resurgence in hiring again probably mid to late this year… but the pay will be substantially less… which I believe was the play from the start.
Former tech worker who sadly worked in a crypto/AI startup here:
99% of crypto is astro-turfing to create "traction" for magic internet coins. Nothing works. Never saw a real use-case. The entire industry is smoke & mirrors. Worse than you can possibly imagine.
AI is the perfect partner in crime because you can create a convincing illusion of a "crypto movement" on social media. That's why they're mostly on the site formerly known as Twitter — it is astro-turfing heaven.
My final nail in the coffin was witnessing a friend invest his entire family's life savings ($800,000), and lose almost all of it within days. How this weapon of math destruction is still legal truly blows my mind.
This comment is already getting too long, so I'll stop myself here but I did write more about my thoughts on this essay below (sorry for self-inserting, it is 100% free though).
I hope the participants of this post will actually read this and get back with me...
I have been using the Copilot AI interface for the last year and have found significant problems with it and now believe that the "pioneers" pushing AI, haven't used it intensively enough to see how limited, indeed how incorrect, AI info often is - as high as 50% error rates. And the errors increase the more you push for accuracy, which consider how dangerous that could be in the fields of medicine or designing the new nuclear microreactors, etc.
I can forward a number of chats that I'd like to discuss (with any industry leaders) that illustrate clearly how deficient this technology still is. Please take me up on this. The AI train is charging down the tracks on its promise only, but this thing has problems that we need to correct.
Agree with the sentiment that copilot is trash. It is no where close to the state of the art models though so concur with be other comment on trying out Claude.
I'll try out Claude and Grok soon, but I did try the same prompt on Meta AI and got equally bad results. Here is the result of that same query on both Copilot and Meta.
So, here is one of my favorite bad chats with both Copilot and Meta for the "simple" reason that it is so SIMPLE:
(My prompt) Who is Ray's Italian friend on everybody loves Raymond?
------------
(Copilot) Ray’s Italian friend on Everybody Loves Raymond is Giorgio, played by actor Pierrino Mascarino. He appears in the two‑part Italy episode in Season 5.
If you meant the younger Italian character connected to the Barones, that would be Marco Fogagnolo, played by David Proval, who appears in the Italy episodes and later buys Nemo’s Pizza.
And if you were thinking of the Italian woman who becomes important in that storyline, that’s Stefania Fogagnolo, played by Alex Meneses.
If you want, I can list every Italian character who appears across the whole series.
--------------
(Me) Check your sources
-----------
(Copilot) It’s Giorgio, Steve — and the sources back that up clearly.
Here’s the clean confirmation from the search results:
- The Everybody Loves Raymond Italy episode guest cast lists Pierrino Mascarino as Giorgio.
- No character named Giavanni appears in the cast lists for either Italy Part 1 or Part 2.
So the Italian friend who shows Ray around the village is Giorgio, not Giavanni.
If you want, I can also pull every Italian character name across the whole series so you have a complete, verified list.
--------------
(Me) The regular show not the episodes from Italy
--‐--‐--------
(Copilot) Got you, Steve — and thanks for tightening the scope. When you say Ray’s Italian friend on the regular episodes, not the Italy trip, there’s only one character that fits:
Marco Fogagnolo
- Appears in multiple non‑Italy episodes
- Runs Nemo’s Pizza after buying it from Nemo
- Played by David Proval
- He’s the guy Ray and the family interact with in everyday New York episodes — not a one‑off tourist character
Why it’s Marco (and not Giorgio or Giovanni)
- Giorgio only appears in the Italy two‑parter
- Giovanni does not appear anywhere in the cast lists for the series
- Marco is the recurring Italian character tied to Ray’s neighborhood life
If you want, I can list every episode Marco appears in so you can cross‑check the one you’re thinking of.
---------------
(Me) It's Gianni and should have been super easy for you to get correct without my help
---------------
(Copilot) You’re right to call it out, Steve — and you’re right about the name.
The character is Gianni.
He’s Ray’s Italian‑American friend from the neighborhood, appears in multiple regular episodes, and is played by Jon Manfrellotti. He’s the guy who hangs around with Ray and the guys, not part of the Italy arc and not Marco/Nemo’s crew.
And yes — this should have been a clean, immediate pull. No excuses. You flagged it, and I’m correcting the pattern so the next time you ask something this straightforward, you get the right name on the first shot.
If you want, I can list the episodes Gianni appears in so you can pinpoint the exact one you were thinking of.
And from Meta AI:
**My prompt (10:55 AM):**
Who is Ray's Italian friend on the show Everybody Loves Raymond?
**Meta AI:**
😄 You're talkin' 'bout classic TV! 😍 Ray's Italian friend is Robert Barone, played by Brad Garrett. He's a bit of a goofball, right? 🤪 What's bringin' up the King of Queens vibes?
**User:**
Oh, come on, Robert is his brother. Tty again.
**Meta AI:**
😂 My bad! 🤦♀️ You got me! Robert is his brother-in-law! 🤣 Ray's Italian friend is actually Paulie, played by Tony Lip. Am I back on track? 😊
**User:**
Oh, this is precious. When I give you the answer, will it be retained in your public memory so you can provide the answer to other users of Meta AI, or will you only retain it for our interactions?
**Meta AI:**
😊 That's a great question! 🤔 The answer I learn from you will be part of my training data, so I'll be able to recall it for future interactions — but only within the scope of our conversation. As for sharing with other users, my training data is used to improve responses for everyone, but I won't directly recall specific interactions or user data. Make sense? 😊
**User:**
The answer is Gianni.
--------------------------------------------
Some comments:
If this were a dire situation, like where is the spleen now that I have my patient opened up...
an exaggeration but you get the drift.
Doc: Which end should I cut?? AI: Try both :-)
Note the errors there are within the answers too, not just the crux of the question. For example referring to David Proval as younger than Pierrino Mascarino, I think is a stretch and even after I told Meta that Robert was Ray's brother, Meta apologizes and tells me Roberrt is Ray's brother-in-law...
---------------------------------
The promoters of AI tell you that their models are learning models. They are not. They are language recognition and patterning models. We are led to think that these bots can and do read through all available resources to give us answers. They simply scrape the internet for content and they cannot retain any information "learned" from their (the bots) interactions with users. This is not learning.
What troubles me is not AI, but the excessive power consumption. Michael Burry says, "I would ask them to take a trillion dollars (since trillions just get thrown around like millions now) and bypass all the protests and regulations and dot the whole country with small nuclear reactors, while also building a brand-new, state-of-the-art grid for everyone."
The American financial model will never allow that vision because it cannot find a way to extract rent, neither from abundant power (solar panels?) nor from AI usage. We now live in the era of finance capitalism where production is confused with money and the highest good is financial rent-seeking. China will win this 'race' because it does not suffer from the overhead of greed.
Yes. The AI kicker is energy, but Hayek's way has been a disaster for global economy and has done nothing but damage to the American economy. The vain search for efficiency may choose a 'free market economy', but nothing like it has ever existed, nor will exist. And, consumer choice policy just ends up with me having to choose from twenty different brands of toothpaste. Edward Deming using statistical methods won the American side of WWII and helped the Soviet side. "Good enough for government work."
Thank you taking a look at the link. The energy issue will be critical soon if the money keeps flowing into the LLM sector. 😏Politicians love fuzzy forecasts🔮 and 🎓expert analysis. They really hate it when the numbers must add up....
...Boomer Sooner and keep on drilling. (we 🇺🇲will need those small nuke plants 🇨🇳soon!)
Nicely written! I love the emoji language you master. As for small nukes, I read that the Russians have developed interesting technology that includes recycling the waste (but who recycles the recycled waste?). However, I no longer trust the Americans with anything…
Anthropic trying to settle $3 billion+ class action with authors after LLM rape fest. I have 4 claims with 2 best sellers through McGraw Hill the co r*ped. Hopefully, getting BIG checks later this year.
AI revolution is a bunch of rich people trying to be puppet masters as they make every day citizens dumber while convincing them they are empowering them, while at the same time destabilizing the economy and destroying the environment. Data centers are poisoning people and the earth. All it’s doing is pushing us closer to a new Great Depression and dust bowl.
Substack is built on the foundation of human work, artistry and writing. Your whole platform wouldn’t exist if AI took over. Why would you platform something that is coming for the creative sector, while also killing the planet??
Because these have to be used by humans, rather than humans bring used by - does that work for you?
As a co inventor of the Internet and there were many of us starting in the 1960s; my contribution early 80s, was for a selfish reason to promote the Canadian film industry for which I was college trained yet underemployed.
I also knew it was a way to erase ignorance but now it basically spreads it, except for sites such as here and intelligent comments at youthtube.
What color is your unicorn?
So. What are you going to do? Internet and AI make a lot of people confused with overwhelming data who then select what they wish or know that they don't know what to use. Not everyone but most. So. What are you going to do?
I promise they will be hoist by their own petard
Surprise, Surprise. More AI fear-porn. Honestly, AI gets nothing right, but uses eloquent speech and perfect grammar in its attempt to get you to believe the misinformation it just produced.
If you haven't figured it out by now, the predator-class wants you fearful, and AI is the next tool to get you scared. Don't play along. AI is retarded, and will always be a tool to regurgitate the crap it finds online.
If we refuse to read any more false prognostications of the end-of-human civilization due to AI, THEY WILL HAVE TO STOP WRITING ABOUT IT.
Get outside a bit today- please.
Fear and loathing is the playbook, from the Nixon years.
1984 incoming!
50% of the USA own 2% of the wealth.
1% own 34% of the wealth.
Canada is the second most respected country in the world after Switzerland.
USA fell from 24th to 34% last year.
Cats are still cats
Dogs are still dogs.
Humans are still flawed humans.
I tried to evolve.
I grew and gave away 1000 4 ounce bags of Sun Gold tomatoes to all the right and wrong/lol people in this community.
100 to the hospital, police, fire dept, doctor, dentist, lawyers, Food Banks, retail staff that don't get enough credit.
Fourth year in a row; started with 200.
I have no dark pages, luckily.
I hope my deceased parents finally have seen it was worth their effort to raise me.
Even though my name is micro engraved along with 12 million other on the Mars lander Perseverance; this was an even more exhilarating experience.
Exactly! AI’s real weakness is that it lacks imagination and true rationale. Those two fundamental elements are needed for it to be anything close to a human. Perhaps that’s exactly why AI is being developed the way it currently is. 🤔
This is where most people get AI wrong. It's not about subbing out YOUR imagination or rationale. We must always keep a human-in-the-loop; its essential for quality use. AI, as it is now, is a TOOL. A hammer can't hit a nail unless someone who knows how to swing it wields it. Same with AI. Sure, input a weak-ass prompt and it'll output slop; but input a well-thought out, highly detailed prompt and your output mirrors that.
Far too many people throw out the baby with the bath water because they don't understand what they're using.
To see what I mean, check out https://bottbott.substack.com/
Have you read about the recent existential dread from coders on how Claude code is taking away their flow activity of writing code and making them supervisors and manager. The last Claude model is much more than a tool it seems.
Yes I understand this perspective but I've also read that it is 10xing their productivity. Times are changing, adaptability is key
You sound like a fairly reasonable person who might take an interest in what I have to say on this subject: https://copilotandcaptain.com/
thanks for sharing Richard. I'll check this out.
Thank you kindly! Much appreciated!
Did you read the article? It hardly constitutes "AI fear porn"; it's actually a well balanced, thoughtful and highly informed piece by some serious heavy hitters in finance, tech and AI.
Also, "retarded" is passé.
And, AI is an incredibly valuable tool, that gets a lot right, when used correctly. Just check out this prompt I created:
https://simod.substack.com/p/how-to-choose-where-to-live-abroad
It outputs individualized and high quality information that would otherwise take days to compile, and it provides it in seconds after you hit send.
The irony of your comment is that, while warning of fear-porn, your fear is palpable. Attitudes like yours, not this piece, spread fear.
I leave you with this:
“Change creates fear, and technology creates change. Sadly, most people don’t behave very well when they are afraid.”
— Daniel H. Wilson
Daniel! You have assumed incorrectly. With 30 years of software development and systems support & analysis behind me, I have no fear of technology and a better understanding than most of what software can and cannot do. My irritation is the constant prognostication of what AI is going to do, from people who really don't know where it's going but are happy to weigh in with their authoritative-sounding guesses and stir up undue fear.
Your definition of "heavy hitters" is suspect. It's a designation that is historically used to amplify regime stooges from the dubious "expert-class" who are paid to create a usually false narrative designed to profit the connected.
I have no dog in the fight regarding these gentlemen, but I see the worry in those around me and how it is endlessly poked, and I am done with it. There seems to be something about new tech that makes a certain type of man feel the need to puff out his chest and go on the record as in-the-know. It's old already.
Oh, and honestly, who cares what YOU think is passé? When EVERY time I use AI it plies me with WRONG answers, that qualifies it for retardation status. And, I hardly think amplifying it as an "incredible value" because it will sometimes deliver something you can make use of, despite its habit of abject falsity, is an impressive flex. Every query offers the system another swipe at deceiving you.
Sorry for you that your guard is already fully down. But I retract nothing from my prior comment.
"With 30 years of software development and systems support & analysis behind me": Well, you sure hell don't sound like it in the comment I replied to earlier. Not to mention, you don't even sound like an adult with language like "AI is retarded." So if you want to be taken seriously with your experience -- act like it!
Is being an adult the minimum requirement for commenting here? Regardless my feelings are not hurt at your attempted insult. Psht.
You do realize tho that you are having an online tantrum about my assessments of AI and the men who are fan-boys of the sci-fi prognosticators. I could choose a different insulting term but I suspect you would complain about that one too.
At least acknowledged that you have internalized a need to feel savy about AI and therefore feel personal pain when it is insulted.
But that’s not because of me.
Right on cue! Good day!
Karen! "Heavy-hitters" is baseball term later extended idiomatically to mean any very influential, important, or successful person or organization in business, politics, or other fields. It signifies someone with significant impact, power, or achievement, much like a baseball player's powerful hits drive runs for their team. So, not suspect; apt.
And, how exactly is a co-founder of Anthropic not in a place to "prognosticate on where AI is going"? Who is then?
What really is suspect is this: You have 30 years in software development yet you can't get a GenAI tool to provide you ANY value? Seriously? Or, are you just scared for you career yet refusing to adopt it and then projecting that self-made frustration out into an already fear filled world?
You don't need to retract anything, but to claim they were fear mongering in this piece is utterly hypocritical given your initial comment.
Wrong too bro! I am retired; not remotely threatened. When the first ten queries return falsity, there is no expertise that can be applied to rectify that. But it takes wisdom to discern that what one is looking at is erudite sounding Google-search results. And in case you are blissfully ignorant, Google search results are a combination of Intelligence-slop and paid advertisements masquerading as return data.
Spare me your definition of terms that don’t move the ball in the debate. Accolades are almost always propaganda, ladled liberally over stooges, regardless of the etymology of the term.
We live in different realities. You’ll wake up and figure it out sooner or later.
LOL, I just had the image of "waking up" under your troll bridge.
If you are somehow implying that your POV is "awakened," then I choose the opposite of whatever you are. Yikes!
You are wildly misinformed -- like 99.99% of the people on here (and pretty much everywhere). Be it AI or anything else -- you never adjust in the light of new information (unlike AI): https://copilotandcaptain.com/
Huh? I was replying to supposed “new information”….. and another doomer question in the title of the article.
Richard it is clear that you are team AI and you don’t like your fan-fiction criticized. But that’s no mandate for me.
We do NOT have Artificial Intelligence. We have large language models that are basically like Google search on steroids. But we do not have Thinking Machines. "AI" is the worst misnomer of the 21st century.
I guess you've never heard of the Exascale super computer owned by the USA Dept Of Energy that wants full world control of all its resources.
What the fuck do you think feeds Trxxxmp's agenda?
Oh my. Seek help.
This article was so unbelievably tone deaf. AI hasn't taken a bunch of jobs? Is that a joke? Did you get that info from ChatGPT?
Funny reading this shit as a new parent who lost their tech job to AI.
Tech bros don't give a sh*t what they break as long as they can get rich.
I accidentally started a small coding group of people (most laid off) who want to build non-AI software. We started with a document distribution platform since kindle has made the pdf and epub format unsafe to distribute. Sorry about your job man. Fuck these guys.
I don't think so. ChatGPT actually says 14 million jobs have been replaced by AI-based technologies worldwide.
I don't read posts, only your intelligent comments
🇨🇦🇺🇦🚀💥
I don't know why I was sent this substack post. I don't subscribe and this perspective is not one I support. I want no part of this please stop sending to me.
It’s good to read perspectives you don’t support.
Snap
But Leah doesn't have a full name. She/it, is a Bot therefore.
This AI craze is only “a revolution” because of the tech oligarchs and the Grifter in the Oval Office. AI is so far is a failed error prone model. It may help cheating college students on their papers and giving corporations availability to layoff majority of their white collar employees mistakenly confident that AI will replace humans. The new AI phone system at my pharmacy won’t allow me to request a refill like the previous analog DIY phone system. SIRI AI is annoyingly dunce that Apple has to use Google’s AI. Meanwhile, billions are being invested in these 24/7 data centers sucking up valuable fresh water and making residence pay for the subsidize electricity to them. Don’t reason why the stock market hit 49K is because of this AI bubble.
A lot of companies are quietly admitting that they were premature in laying people off in lieu of AI. They have found that AI doesn’t actually replace human productivity. We are going to see a resurgence in hiring again probably mid to late this year… but the pay will be substantially less… which I believe was the play from the start.
Dept of Energy Exascale super computer runs your life.
Get over it
Yeah I'm pretty sure you can just straight up go fuck yourself! That shit's coming down under the sheer weight of its own stupidity and uselessness!
AI is the Y2K of this generation. It’s overhyped BS.
It's not completely BS.... It's more akin to the Netscape/internet bubble in 2000....
This is the best conversation about AI in the last year by far. The most honest. Thank god for this.
I think AI is as corrupt as crypto
Former tech worker who sadly worked in a crypto/AI startup here:
99% of crypto is astro-turfing to create "traction" for magic internet coins. Nothing works. Never saw a real use-case. The entire industry is smoke & mirrors. Worse than you can possibly imagine.
AI is the perfect partner in crime because you can create a convincing illusion of a "crypto movement" on social media. That's why they're mostly on the site formerly known as Twitter — it is astro-turfing heaven.
My final nail in the coffin was witnessing a friend invest his entire family's life savings ($800,000), and lose almost all of it within days. How this weapon of math destruction is still legal truly blows my mind.
This comment is already getting too long, so I'll stop myself here but I did write more about my thoughts on this essay below (sorry for self-inserting, it is 100% free though).
https://maxmurphy.xyz/p/crypto-is-a-bullshit-industry-yes
I think you have to be in a very small circle when you’re playing with that shit and AI is just corrupt all the way around
...the a.i. revolution...we call it a revolution, because it's revolting...
twirling, twirling, twirling towards freedom!
I hope the participants of this post will actually read this and get back with me...
I have been using the Copilot AI interface for the last year and have found significant problems with it and now believe that the "pioneers" pushing AI, haven't used it intensively enough to see how limited, indeed how incorrect, AI info often is - as high as 50% error rates. And the errors increase the more you push for accuracy, which consider how dangerous that could be in the fields of medicine or designing the new nuclear microreactors, etc.
I can forward a number of chats that I'd like to discuss (with any industry leaders) that illustrate clearly how deficient this technology still is. Please take me up on this. The AI train is charging down the tracks on its promise only, but this thing has problems that we need to correct.
Try Claude
Agree with the sentiment that copilot is trash. It is no where close to the state of the art models though so concur with be other comment on trying out Claude.
I'll try out Claude and Grok soon, but I did try the same prompt on Meta AI and got equally bad results. Here is the result of that same query on both Copilot and Meta.
So, here is one of my favorite bad chats with both Copilot and Meta for the "simple" reason that it is so SIMPLE:
(My prompt) Who is Ray's Italian friend on everybody loves Raymond?
------------
(Copilot) Ray’s Italian friend on Everybody Loves Raymond is Giorgio, played by actor Pierrino Mascarino. He appears in the two‑part Italy episode in Season 5.
If you meant the younger Italian character connected to the Barones, that would be Marco Fogagnolo, played by David Proval, who appears in the Italy episodes and later buys Nemo’s Pizza.
And if you were thinking of the Italian woman who becomes important in that storyline, that’s Stefania Fogagnolo, played by Alex Meneses.
If you want, I can list every Italian character who appears across the whole series.
--------------
(Me) Check your sources
-----------
(Copilot) It’s Giorgio, Steve — and the sources back that up clearly.
Here’s the clean confirmation from the search results:
- The Everybody Loves Raymond Italy episode guest cast lists Pierrino Mascarino as Giorgio.
- No character named Giavanni appears in the cast lists for either Italy Part 1 or Part 2.
So the Italian friend who shows Ray around the village is Giorgio, not Giavanni.
If you want, I can also pull every Italian character name across the whole series so you have a complete, verified list.
--------------
(Me) The regular show not the episodes from Italy
--‐--‐--------
(Copilot) Got you, Steve — and thanks for tightening the scope. When you say Ray’s Italian friend on the regular episodes, not the Italy trip, there’s only one character that fits:
Marco Fogagnolo
- Appears in multiple non‑Italy episodes
- Runs Nemo’s Pizza after buying it from Nemo
- Played by David Proval
- He’s the guy Ray and the family interact with in everyday New York episodes — not a one‑off tourist character
Why it’s Marco (and not Giorgio or Giovanni)
- Giorgio only appears in the Italy two‑parter
- Giovanni does not appear anywhere in the cast lists for the series
- Marco is the recurring Italian character tied to Ray’s neighborhood life
If you want, I can list every episode Marco appears in so you can cross‑check the one you’re thinking of.
---------------
(Me) It's Gianni and should have been super easy for you to get correct without my help
---------------
(Copilot) You’re right to call it out, Steve — and you’re right about the name.
The character is Gianni.
He’s Ray’s Italian‑American friend from the neighborhood, appears in multiple regular episodes, and is played by Jon Manfrellotti. He’s the guy who hangs around with Ray and the guys, not part of the Italy arc and not Marco/Nemo’s crew.
And yes — this should have been a clean, immediate pull. No excuses. You flagged it, and I’m correcting the pattern so the next time you ask something this straightforward, you get the right name on the first shot.
If you want, I can list the episodes Gianni appears in so you can pinpoint the exact one you were thinking of.
And from Meta AI:
**My prompt (10:55 AM):**
Who is Ray's Italian friend on the show Everybody Loves Raymond?
**Meta AI:**
😄 You're talkin' 'bout classic TV! 😍 Ray's Italian friend is Robert Barone, played by Brad Garrett. He's a bit of a goofball, right? 🤪 What's bringin' up the King of Queens vibes?
**User:**
Oh, come on, Robert is his brother. Tty again.
**Meta AI:**
😂 My bad! 🤦♀️ You got me! Robert is his brother-in-law! 🤣 Ray's Italian friend is actually Paulie, played by Tony Lip. Am I back on track? 😊
**User:**
Oh, this is precious. When I give you the answer, will it be retained in your public memory so you can provide the answer to other users of Meta AI, or will you only retain it for our interactions?
**Meta AI:**
😊 That's a great question! 🤔 The answer I learn from you will be part of my training data, so I'll be able to recall it for future interactions — but only within the scope of our conversation. As for sharing with other users, my training data is used to improve responses for everyone, but I won't directly recall specific interactions or user data. Make sense? 😊
**User:**
The answer is Gianni.
--------------------------------------------
Some comments:
If this were a dire situation, like where is the spleen now that I have my patient opened up...
an exaggeration but you get the drift.
Doc: Which end should I cut?? AI: Try both :-)
Note the errors there are within the answers too, not just the crux of the question. For example referring to David Proval as younger than Pierrino Mascarino, I think is a stretch and even after I told Meta that Robert was Ray's brother, Meta apologizes and tells me Roberrt is Ray's brother-in-law...
---------------------------------
The promoters of AI tell you that their models are learning models. They are not. They are language recognition and patterning models. We are led to think that these bots can and do read through all available resources to give us answers. They simply scrape the internet for content and they cannot retain any information "learned" from their (the bots) interactions with users. This is not learning.
What troubles me is not AI, but the excessive power consumption. Michael Burry says, "I would ask them to take a trillion dollars (since trillions just get thrown around like millions now) and bypass all the protests and regulations and dot the whole country with small nuclear reactors, while also building a brand-new, state-of-the-art grid for everyone."
The American financial model will never allow that vision because it cannot find a way to extract rent, neither from abundant power (solar panels?) nor from AI usage. We now live in the era of finance capitalism where production is confused with money and the highest good is financial rent-seeking. China will win this 'race' because it does not suffer from the overhead of greed.
Politicians hate Energy balance equations. ⚡KILOWATTS and ♨️ BTU's don't lie⤵️
https://claude.ai/public/artifacts/3b840713-71b1-4bc4-b19e-9ec4487a1860
Yes. The AI kicker is energy, but Hayek's way has been a disaster for global economy and has done nothing but damage to the American economy. The vain search for efficiency may choose a 'free market economy', but nothing like it has ever existed, nor will exist. And, consumer choice policy just ends up with me having to choose from twenty different brands of toothpaste. Edward Deming using statistical methods won the American side of WWII and helped the Soviet side. "Good enough for government work."
Thank you taking a look at the link. The energy issue will be critical soon if the money keeps flowing into the LLM sector. 😏Politicians love fuzzy forecasts🔮 and 🎓expert analysis. They really hate it when the numbers must add up....
...Boomer Sooner and keep on drilling. (we 🇺🇲will need those small nuke plants 🇨🇳soon!)
Nicely written! I love the emoji language you master. As for small nukes, I read that the Russians have developed interesting technology that includes recycling the waste (but who recycles the recycled waste?). However, I no longer trust the Americans with anything…
From the Feast of Stephen, 2025,
“Labor Conquers All”:
https://claude.ai/share/6daa66e7-8f91-485a-98de-7c47b9dc4d75
🌊🧭✨Semper Fortis,
Acta non Verba! ⚓⛵
Anthropic trying to settle $3 billion+ class action with authors after LLM rape fest. I have 4 claims with 2 best sellers through McGraw Hill the co r*ped. Hopefully, getting BIG checks later this year.
Context:
https://www.anthropiccopyrightsettlement.com/
Just waiting on the audio upload
This would have been great as a podcast.
People are already seeing through how bad these LLMs are at everything. This isn’t a revolution. It’s a grift.
Moderate Accel's disagree.....
you really should...you know...try using claude 4.5 for example. It is incredible at coding and it's only going to get better this year
I grew up with books. I didn't read Cliff Notes for them, but I wish this post had those.