@CardGeek said:
Why is machine learning important? What do you think these systems are going to accomplish? It sounds like the pay off is putting people out of work.
In 1999 a lot of the companies popping up were being run by 20 year olds. In 1999 a lot of the business popping up had no idea how they were going to make money off of whatever it was they were making. Ironically many of the ones who survived ended up making money off annoying popups.
It’s not about replacing people but anytime something can outperform a person it’s time for people to move aside and focus on another task.
We're having a hard time figuring out what these other tasks are. Maybe someone should design an AI to figure out what other tasks humans should be doing.
Most of the time when I get an AI on the phone, I'd rather be talking to a human.
At least one major US Bank has invested approx 20 Billion into AI/Machine Learning in the last 12 years , yep prior to it even being "fashionable" to do so. They would not spent that kind of coin if getting rid of employees was not at minimum their secondary objective.
It's the singer not the song - Peter Townshend (1972) Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
@CardGeek said:
We're having a hard time figuring out what these other tasks are. Maybe someone should design an AI to figure out what other tasks humans should be doing.
Most of the time when I get an AI on the phone, I'd rather be talking to a human.
Support bots are hardly AI. Most are script based and have nothing AI related.
In 2023, this was the definition of AI that the White House used while drafting it's executive order on AI:
b) The term “artificial intelligence” or “AI” has the meaning set forth in 15 U.S.C. 9401(3): a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. Artificial intelligence systems use machine- and human-based inputs to perceive real and virtual environments; abstract such perceptions into models through analysis in an automated manner; and use model inference to formulate options for information or action.
Sometimes phone support bots decide to hang up on me.
I don't know if you remember Ask Jeeves? Search engine in the late 90s that allowed you to ask questions in plain English. Parsed English based on a language model to figure out what you were asking.
All of these AIs are following a script. It is a more complex script than a phone support bot's script. That's just how computers work. Stuff like the ChatGPT is using cashed websites and spitting out a formatted mashup.
None of these AIs are doing anything creative. They can write a song. Writing a song is just playing notes though. Anything can be a song. It's original as long as it doesn't match anything in the database. When you think about it, it makes being a human seem simple and stupid. Computers have been able to write songs since the 80s. Maybe even earlier.
I think most of the AI news out there is completely fabricated. If there is something amazing out there I haven't heard of it yet.
At least one major US Bank has invested approx 20 Billion into AI/Machine Learning in the last 12 years , yep prior to it even being "fashionable" to do so. They would not spent that kind of coin if getting rid of employees was not at minimum their secondary objective.
The stage coach driver union agrees with this message.
@1982FBWaxMemories said:
^ overly simplistic comment especially from the only one here who seemed to know the meaning of "japery" without having to look it up.
Well. It’s a long and nuanced discussion and I don’t have all the answers and this isn’t the environment. I wasn’t meaning to be trite.
Well AI is an umbrella term. At best chat bots are using RNNs. Fairly antiquated.
In 2023, this was the definition of AI that the White House used while drafting it's executive order on AI:
b) The term “artificial intelligence” or “AI” has the meaning set forth in 15 U.S.C. 9401(3): a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. Artificial intelligence systems use machine- and human-based inputs to perceive real and virtual environments; abstract such perceptions into models through analysis in an automated manner; and use model inference to formulate options for information or action.
I dont have any problem with that description for that context. It makes as much sense as it needs to.
Sometimes phone support bots decide to hang up on me.
I’m sorry that happened to you. You didn’t deserve that.
I don't know if you remember Ask Jeeves? Search engine in the late 90s that allowed you to ask questions in plain English. Parsed English based on a language model to figure out what you were asking.
Ask Jeeves was not using a language model as would be considered today. It was a keyword tokenizer.
All of these AIs are following a script. It is a more complex script than a phone support bot's script. That's just how computers work. Stuff like the ChatGPT is using cashed websites and spitting out a formatted mashup.
So that’s not how LLMs work and you’re conflating that with the synopsis they provide based on what they are able to interpret.
None of these AIs are doing anything creative. They can write a song. Writing a song is just playing notes though. Anything can be a song. It's original as long as it doesn't match anything in the database. When you think about it, it makes being a human seem simple and stupid. Computers have been able to write songs since the 80s. Maybe even earlier.
You’re again focused on LLMs which are a very specific implementation of a neural network.
I think most of the AI news out there is completely fabricated. If there is something amazing out there I haven't heard of it yet.
You’re speaking at a very high level on a wide ranging topic. There is a lot of BS out there but the only thing you’re touching on is LLMs for customer service. I’ll agree that’s boring.
@CardGeek said:
You're attempting to make this sound a lot more complex than it actually is.
What am I trying to make sound too complex? We’re not discussing anything specific enough to be complex. It’s such a huge topic and we’re barely past “AI”. For 30 years the best we had was an encoder-decoder architecture and it was right to mock AI as merely programmed response however complex the pipelines got. Transformers and the introduction of context has unlocked NLPs to become smart enough to hang up on you… and perform many other complex tasks which people have struggled with. While automation gets the most news the real value, I think, comes from the ability of the convolutional layers to deduce the most effective correlations across remarkably large datasets without requiring brute force compute.
They're not being smart when they hang up. That's just what they were scripted, programmed, to do in that situation.
Computers are really good at going 1101000100100101010001001010111100101010101001 really really fast. That's really all they know how to do. Somebody tells them to do something when they a particular string.
Artificial Intelligence is an interesting term. Artificial, fake, Intelligence. It's a lot of smoke and mirrors.
They've been brute forcing for 40 or more years now. Some of these companies have brute forced the entire internet.
Humans having real intelligence isn’t a meaningful advantage. Our advantage is consciousness. Intelligence is a measurement of the rate of knowledge acquisition. In that sense, compared to many humans, machine learning is stupid as it lacks enough memory to maintain coherence. But we don’t need a general knowledge solution to make use of machine learning. And it’s much more than beep boop here’s a stream of bits. Don’t pretend you know what you’re talking about. Moral of the story is… don’t worry about buzzwords. If you’re hung up on how smart it is you’ve missed the point 10 stops back.
Don't you love it when someone uses the low (April 5) to gage returns when ydt is actually around half that? Good discussion though. My 2 cents.... data would indicate that equities are overvalued. Historically, after 3 years of double digit returns it gets a little dicey. Fed cutting rates bodes well for short term gains. Inflation kind of under control. Jobs market and consumer sentiment probably mystery drivers. Keep it simple and diversified. Time in always beats timing. Sport cards have zero bearing on market performance.
@sayheywyo said:
Don't you love it when someone uses the low (April 5) to gage returns when ydt is actually around half that? Good discussion though. My 2 cents.... data would indicate that equities are overvalued. Historically, after 3 years of double digit returns it gets a little dicey. Fed cutting rates bodes well for short term gains. Inflation kind of under control. Jobs market and consumer sentiment probably mystery drivers. Keep it simple and diversified. Time in always beats timing. Sport cards have zero bearing on market performance.
I agree with this. It's easy to get carried away in forums but collector's items such as cards and artwork are always downstream of stocks and the broader economy. They are generally correlated with crypto as well (with a couple exceptions, I think of most cryptocurrencies as "collectibles"). This framework is helpful for understanding these market as reservoirs for when excess liquidity from the broader economy spills over.
The lone exception to this thought framework was Q3 2020 - Q1 2021. That was a one-time boost to this market that transcended historical patterns and brought new entrants back in. As someone who was selling unopened I'll never forget it. I started buying back in after National in summer 2023. The only way you get a true feel for the market is to participate. Generally speaking I felt I was much too early to buy throughout '23 and 2024. In crypto some call this a PVP (player vs. player) market. You are trading against established market participants and cannot count on new entrants as exit liquidity. Only in 2025 did the collectibles market get stronger which I would attribute to broader wealth effects that result from stock market at ATHs. Easy to identify with precious metals and crypto also at ATHs.
The main headwind is still lack of new entrants. Big questions - what few new entrants come in - where will they spend? And do scarce conditions exist in that market? I continue to believe there will not be a catalyst to bring new entrants to the hobby other than those who are "aging" back into it (most hobby entrants/returning story goes that they hit 40 and reached a level of wealth that allowed more discretionary spending). Saw good price growth in 1996-2000's unopened, MTG and Pokemon (late millennial products). That tracks with the demographic of those "aging" in to the market.
Despite the improving market, as an investor you still have the ongoing challenge of choosing the right product (player, or year/brand in unopened). Some of my unopened stuff has doubled while admittedly, some of my "investment" unopened from 2023 only kept pace with shipping price increases (inflation). Kind of sad, but it's easy to make mistakes in the hobby. And as I have stated before, you always have what I've often referred to as the 20% hurdle when liquidating. Generally this is 13-15% eBay fees, 3-5% shipping cost, 3-5% slippage (returns, theft, act of god). And when all is said and done, capital gains tax.
At the end of the day I still do this for fun/completing collecting goals, though there is an investment mindset behind purchases that cannot be ignored.
@CardGeek said:
Don't pretend to know what I'm talking about? Bro. You're Mr. Buzzword in this conversation.
You should step back and think about what those buzzwords are describing.
Everything in computing is strings of bits. That's all computers do. High or low.
Binary is a representation of data but everything in computers is not high or low and that’s not all computers do. Binary states are foundational for classic computing but that’s like saying humans are just cells and nothing more. If you have a pile of cells well you’re looking at a human. That pile of leaves over there? That’s cells bro. Thats all humans are.
My area of expertise is in vision transformers so if you want to discuss details there that’s what I’m most interested in. It’s also most applicable to what PSA should be investing in.
There is definitely still a lot of confusion out there regarding AI. If you want a little insight into how a LLM works I highly recommend this 30 minute video (one part of a really good series). It's a little dated but I love it and have shared it with several family members who had thought this stuff was magic. It isn't. https://youtu.be/wjZofJX0v4M?si=ON81NJQiRSvCrW1u
@CardGeek said:
Why is machine learning important? What do you think these systems are going to accomplish? It sounds like the pay off is putting people out of work.
I think that is exactly it. Its not as if we are waiting for it to happen either. Its already happened. Ask a kid fresh put of college wanting to get a job as a Jr. Programmer how wonderful the experience of trying to find a job now is. What people are doing with AI right now is finding tasks/jobs that can be done with AI so they can let people go or not hire them. They can spend less not paying employees and their salaries and their fringe benefits and they can rent one less floor of a building. Then the AI can do the work in minutes that may have taken the person days weeks months. So for now saving money on labor and the end result is improving profits and beating the # when they report or just improving year over year on the numbers.
To clarify, I am not saying yay, this is a wonderful thing and lets celebrate its arrival and all the consequences. I do think it has some bad potential for the working individual. But just like a wager, you may despise the Chiefs but do you think they will win. If so, bet the Money Line.
And I think right now is just the tip of the iceberg in terms of seeing how much it can do. I just mention one thing where it can benefit companies in terms of decreasing costs while getting similar or better production. But we had the ChatGPT moment. Which happened with x amount of compute power and older less powerful Nvidia chips I would think. What is going to happen when all these data centers are built and powered? What additional progress will be made?
But I think what is interesting is between now and then so much money will be spent to build these data centers. They dont even have the energy to let all of these planned data centers run now so they are reopening shutdown nuclear plants. The spending impacts so much more than a company like Nvidia who makes the chips because you cant run a data center with an Nvidia chip. Cant just throw it a patch of dirt and say work. Think things. Answer questions. You have to build the building. Construction. Materials. You have to have the servers. You need the equipment, cables, storage. Cooling. Energy. It would be really cool to see the Accounts Payable list of vendors for Microsoft Amazon Meta Oracle Google.
If AI ends up being a bust, and I dont think it will, I dont think we would see it until 2029 2030. The buildout will be enormous and the nuclear power needed to fuel so much of it may not be ready until 2027 2028. Building a data center is like building an MGM Grand too. Cant get that up and running overnight. Until all the building is complete, which may be never or a ridiculous amount of time cause its like building the railroad or the highways, all the purchasing and building can be enough to make some companies really successful for 3 years or so. Then reevaluate and maybe its time to sell but it could have been a nice ride.
@bgr said:
On one hand... everyone is talking about AI these days. I don't encounter many people who understand what any of it is and what it does. In that sense this is very much a bubble. AI Mania.
I think that is part of the issue too. I think a lot of the people who dont fully grasp it say, I dont get it. Its a bubble. But to really have that opinion that people want to listen to and trust, I think they should get it more and then say I fully understand it and I think its a bubble. I'm not saying people here. Say whatever you want. But more people who are analysts who invest in the financials and healthcare and consumer staples and that is their wheelhouse but then they have a strong opinion on tech. I think its popular to call it a bubble and its getting trendy.
I will admit some people knee deep in tech still say it. Like Sam Altman. But maybe what they feel the value of his company is is overinflated and therefore sure, why shouldn't he think its a bubble. If his company was public possibly the most bubbly stock out there.
I do listen to the guys who are knee deep in tech stocks who specialize in that area and understand it well and most are enthusiastic and dont feel its a bubble. I think its the other guys who are less involved, more diversified with less tech in the portfolio who say that its a bubble.
@bgr said:
My area of expertise is in vision transformers so if you want to discuss details there that’s what I’m most interested in. It’s also most applicable to what PSA should be investing in.
>
>
High and low is actually all computers do. You're typing a lot but you're not saying very much. You keep on saying you want to talk about it.
@UlyssesExtravaganza said:
I think that is exactly it. Its not as if we are waiting for it to happen either. Its already happened. Ask a kid fresh put of college wanting to get a job as a Jr. Programmer how wonderful the experience of trying to find a job now is.
I don't know if anyone else has noticed this, but, ChatGPT is mostly doing the same thing your search engine does. Except, instead of serving you a link to a page with information on it, it's serving the data from the page, mashed up with other data from other pages on the same topic, and typing out that information really slowly, to make it seem like it's more human.
A few years back all these companies started complaining that these AIs were being trained using the companies data. What's happening is, companies like Google, who have been crawling the internet for 40 years or so, have cashed the whole net. They started off making a directory of web sites. But, then they started saving the sites and scraping the databases. For instance, a lot of the time you don't have to click the wiki link because the information from wiki shows up on the right panel of your search results because Google saved the wiki page. They're saying their AI learned it. As if their database is a brain. It's similar to me recording the world series with my DVR and saying my computer remembers watching it. Then distributing the world series and saying my AI is talking about the world series from memory.
@bgr said:
On one hand... everyone is talking about AI these days. I don't encounter many people who understand what any of it is and what it does. In that sense this is very much a bubble. AI Mania.
I think that is part of the issue too. I think a lot of the people who dont fully grasp it say, I dont get it. Its a bubble. But to really have that opinion that people want to listen to and trust, I think they should get it more and then say I fully understand it and I think its a bubble. I'm not saying people here. Say whatever you want. But more people who are analysts who invest in the financials and healthcare and consumer staples and that is their wheelhouse but then they have a strong opinion on tech. I think its popular to call it a bubble and its getting trendy.
I don't disagree. Analysts should be separating the wheat from the chaff. There are companies which are BS and circle-funding to drive valuations. There are companies out there who have no idea what they are doing. We're at the point now where there are tools out there which allow anyone to start labeling and start training so there's a lot of poor models out there that companies are building their valuations off of and these modules are destined for early plateau or complete collapse.
I will admit some people knee deep in tech still say it. Like Sam Altman. But maybe what they feel the value of his company is is overinflated and therefore sure, why shouldn't he think its a bubble. If his company was public possibly the most bubbly stock out there.
I think Altman has embraced his own confidence and he's started to make stuff up (chart-gate)... GPT-5 has regressed for sure. I don't know if it's reasoning collapse or just F1 divergence but they are already raising money... I think like $2T to build the next generation model. We observe that these LLMs tend to get over-fit with relative ease.
I do listen to the guys who are knee deep in tech stocks who specialize in that area and understand it well and most are enthusiastic and dont feel its a bubble. I think its the other guys who are less involved, more diversified with less tech in the portfolio who say that its a bubble.
To the point about jobs, loss, shift, however it shakes out here, there, and everywhere... this technology doesn't do much unless there is economic activity to support, so there's that.
@bgr said:
My area of expertise is in vision transformers so if you want to discuss details there that’s what I’m most interested in. It’s also most applicable to what PSA should be investing in.
>
>
High and low is actually all computers do. You're typing a lot but you're not saying very much. You keep on saying you want to talk about it.
OK. let's do 1 thing at a time.
High and Low is all computers do.
What do computers do.
They take input. They store data. They process information. They produce output.
This is more accurately what computers do.
How do they do it gets us closer to binary and why it's used and so on and so forth.
I started learning about computers in, like, 3rd grade in San Jose, CA. I took my first programming class in 8th grade. I have a strong understanding of databases. I also have a strong understanding of how microprocessors and computers in general operate. I understand the nature of the machine.
That's good. I studied at the University of Minnesota Twin Cities, obtaining two bachelors degrees, one of which was in Computer Science, before attending graduate school at MIT where I also worked in MIT's Lincoln Labs. So sure... feathers out huh.
I don't know if anyone else has noticed this, but, ChatGPT is mostly doing the same thing your search engine does. Except, instead of serving you a link to a page with information on it, it's serving the data from the page, mashed up with other data from other pages on the same topic, and typing out that information really slowly, to make it seem like it's more human.
You're describing the concept of abstraction. The more you squint the less specific something looks so it's reasonable to describe LLMs and Keyword search as similar. They both take input... parse it... process it... and produce outputs. One is just fairly simple and the other produces better output. They work very differently under the hood.
What is this fixation on LLMs. We should all know that they are a NLP which produces a scraped synopsis which they value based on learned interactions. And I won't go past buzzwords here because I can't fit a talk on the concept of Perceptrons into a forum post.
A few years back all these companies started complaining that these AIs were being trained using the companies data. What's happening is, companies like Google, who have been crawling the internet for 40 years or so, have cashed the whole net. They started off making a directory of web sites. But, then they started saving the sites and scraping the databases. For instance, a lot of the time you don't have to click the wiki link because the information from wiki shows up on the right panel of your search results because Google saved the wiki page. They're saying their AI learned it. As if their database is a brain. It's similar to me recording the world series with my DVR and saying my computer remembers watching it. Then distributing the world series and saying my AI is talking about the world series from memory.
It's not a technical statement to refer to a database as a brain. I can't comment on a statement I haven't see or reason why it was made. No one that matters thinks a database is a brain.
The data that companies like Google captured from all those years (20 or so) of user-interaction is priceless. It's labeled data and that's like a pile of money when you're training a model. But what's funny is that when you populate your layers with this data and the proper weights based on that pile of empirical data you start to get a little bit of reason out of your model and as you curate your training data-set you continue to increase the amount of reason. We look at our probability histogram to understand whether we're increasing or decreasing reason. It's much more than a keyword search and a rank-order output function.
My best advice is to question what you think you know. With your background and an open mind I suspect you'll be able to see it more clearly if you take the time to explore. I didn't consider AI much more than a statistical process to filter to the most likely answer for many years, and it wasn't much more than that in practice. We were telling it what to do in the convolutional layers. We were telling it what the important variables were. The big difference really came with the introduction of the transformer architecture.
@CardGeek said:
What are you arguing here exactly?
I suppose, if I was arguing, my position would be that there is more to machine learning that statistical process.
You know, rank-order, that is weight.
With keyword search those weights are static or, at best, periodically updated system-wide. We used to refer to these are hot spots and we would adjust to produce or reduce hot spots in the results.
Rank order is output weight. These weights which are the non-linear part of the data series propagation are input weights which themselves are further perturbed by the recursive redimensioning of the convolutional layers based on the feedforward networks. We encounter this process anytime someone says something that can be interpreted in multiple ways. We think about these possibilities and use the context we have to reform and inform. This is what FFNs and RCLs do. Instead of a keyword search that ranks output they go through each layer to an output and then pass that information back to each layer independently to reinforce learning and produce more accurate output.
I would have thought your advice would have been, "lie".
Thanks for the input/education bgr and others. I will admit I have no hands-on programming tech experience. For my second job out of college I was fortunate to be able to get a Jr. Sales job in Software for a tiny company with about 11 employees. I didn't have a clue but loved the experience and challenge, pre big-internet, of absorbing as much info as possible to be able to have conversations with people and pitch the product.
Second job was more of a reseller of Oracle products. So had to understand just enough about databases to get by even though it was a little more transactional order-taking. But cool to be a part of any product overviews. Back then it was the database licenses being sold and then Developer 2000 Designer 2000 tools to help write applications because in 1997 or so 2000 had such a cool ring to it. Had some other jobs after that in the tech sales space tech training space. Then there was 9/11 and budgets froze and selling was tough and in time I realized I was an average but not outstanding salesman and time to move on.
But my interest in tech companies has remained. I dont have a diversified porfolio. Or if I do, it only looks that way with energy and construction etc. but you still can tie it back to tech with the data center connection. Portfolio is close to 100% tech and has been since about 2018.
So my angle with all of this is mostly business case. I am not able to get down in the weeds. Just looking at the business case and can see that a lot of the companies with the largest market caps believe in AI and cant spend enough which I think for at least a while that means a decent # of companies win. Certainly there are a number of frauds. Just because you say AI 10 times in your earnings call wont make you a winner. Just like Pets.com. Having an internet angle wont make you a winner. Having an AI angle wont make you a winner. So as always, pick right. And there is risk. People dont have to play if they dont want. Folks can stick to value investing. But on the original point just because tech stocks aren't like value stocks, I dont think that should have value investors calling tech a bubble. They are just two different worlds with different rules.
But on the original point just because tech stocks aren't like value stocks, I dont think that should have value investors calling tech a bubble. They are just two different worlds with different rules.
What gives me pause is how much the overall economy appears to be relying on the rampant investment in artificial intelligence. When I look at the model failure rate of LLMs of 80% and general models around 95% that gives me a bit of pause as there's a lot of ROI not being met. A big part of this, from where I sit, is that there's not enough people who know what they're doing to cultivate these models effectively. This is the 'data scientist' role which can only be ill-defined. Maybe it would have been called a statistician more generally when I was coming up - someone who can put data to use using maths. To the point CardGeek was making about being able to make sense of the data for useful purpose - there is a lot of companies flailing around doing nothing useful. There are always frauds... just like dotcom.
But my interest in tech companies has remained. I dont have a diversified porfolio. Or if I do, it only looks that way with energy and construction etc. but you still can tie it back to tech with the data center connection. Portfolio is close to 100% tech and has been since about 2018.
China is doing what China does and putting data centers underwater. Microsoft actually tested this - Project Natick, they determined it was not feasible - many issues - obvious ones if you've seen any movies where people build stuff underwater I guess. There's a lot of cool stuff going on in data-center development by companies like Microsoft, Apple, IBM, Oracle... I think that's a really good area to watch for the pulse of how things are going. I also think that this is going to push development in our energy infrastructure here. If everyone can find a way to go-along and get-along here there's a bright future. Outside of the larger geopolitical brush strokes that is where I would train my eye to see where the success needle is for the "AI Economy".
What's happening is, companies like Google, who have been crawling the internet for 40 years or so, have cashed the whole net. They started off making a directory of web sites. But, then they started saving the sites and scraping the databases.
When I saw 40 years I had to count my rings... I remember talking about BackRub on EFnet #linux... Google was BackRub before it was Googol before it was Google.
One of the most transformational changes to search-algorithm that Google invented was to follow the links with PageRank. Such a simple thing but it had such a major impact on the quality of result they could provide from their search that they won the market.
But I would not say that it is true that Google scrapes databases. They do save sites in Google Index, including raw content that is accessible. I have found that Google's spider does respect robots.txt for Search Result, and noindex for Indexing if done correctly, but this is not the case everywhere and has not always been the case.
@1982FBWaxMemories said:
At least one major US Bank has invested approx 20 Billion into AI/Machine Learning in the last 12 years , yep prior to it even being "fashionable" to do so. They would not spent that kind of coin if getting rid of employees was not at minimum their secondary objective.
At least one major US Bank has invested approx 20 Billion into AI/Machine Learning in the last 12 years , yep prior to it even being "fashionable" to do so. They would not spent that kind of coin if getting rid of employees was not at minimum their secondary objective.
The stage coach driver union agrees with this message.
@1982FBWaxMemories said:
^ overly simplistic comment especially from the only one here who seemed to know the meaning of "japery" without having to look it up.
Well. It’s a long and nuanced discussion and I don’t have all the answers and this isn’t the environment. I wasn’t meaning to be trite.
Agreed it is a long and nuanced discussion BUT the exact example I referenced eliminating jobs was an objective for that firm! They tried to hide their glee but IMHO it's still present
"People shouldn't put their head in the sand. It is going to affect jobs," Dimon told Bloomberg, warning that AI adoption will reshape the workforce. He said automation is already streamlining operations across the bank and added that "there'll be fewer jobs in certain functions," while emphasizing the importance of retraining and redeployment.
CFO Jeremy Barnum said during the bank's annual investor presentation in May that the firm is "asking people to resist headcount growth where possible" while still hiring in "high-certainty areas."
BTW this discussion is a lot more interesting than the self-serving attempt to pump cards as the OP is trying to do! @bgr I'd like to think we are total agreement about that.
It's the singer not the song - Peter Townshend (1972) Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
Have any of you ever read Hitchhiker's Guide to the Galaxy? It's the funniest book I've ever read. The book has been adapted into a Hollywood movie and a BBC TV miniseries. While the BBC series is a better representation of the story, neither of the two do the book justice.
Interestingly, the movie cost over 100 million dollars to make. The BBC series probably cost about 50 bucks and 2 large pizzas to make.
In the book a huge computer is built to calculate the answer to the ultimate question of life, the universe, and everything. The AI tells the scientists that the problem will take 7.5 million years to process and to check back later. Scientists maintain the system for 7.5 million years. After that time the AI tells them that they're not going to like the answer and asks if they really want to hear it. The scientists insist that they want to hear the answer. The AI tells the scientists that the answer is "42".
42, for two, 4 + 2 = 6, six, sex.
You kind of have to read in to it, if you know what I mean.
Sometimes people write a lot that's not really worth reading. Dougles Adams is worth reading.
Here's from the president of Upper Deck on LinkedIn. Maybe a bit of sour apples because, other than hockey and oddball titles, not sure how relevant they are in the market now, but I think this is a fair take:
Everything speculative is up right now: The stock market. Bitcoin. Trading cards…and it's pricing some collectors out.
Here's what the data shows: 50% of consumer spending in the US is coming from 10% of the population. Rich people are carrying the market. Not just in cards, but across every category.
Concert tickets? $1,000 for nosebleeds. Luxury goods? Selling faster than ever. High-end collectibles? Moving at record prices.
When boxes cost $1,000 or more, you're not buying for fun anymore. You're speculating. You're betting that what you buy today will be worth significantly more tomorrow.
And that works. Until it doesn't.
Sports cards have always had speculation built in. You collect a rookie hoping they become a Hall of Famer. But what's happening now is different in scale. And when markets become this speculative, they become fragile.
I'm not saying the sky is falling. I'm saying we need to be honest about what's happening. This isn't sustainable growth driven by collectors who love the hobby. This is wealth concentration driving prices that most people can't participate in.
And if that money suddenly moves elsewhere to bet on something else? We all will feel it.
@craig44 said:
man, a lot of techno talk for a card forum...
That is where that steaming hot cup of Excitement Brew comes from. > @80sOPC said:
Excitement is not brewing at Amazon, at least for the 15-30k corporate employees that are being eliminated.
True. And I heard some discussion this morning, not for the first time, that its not just companies laying off but deciding they dont want to increase head count.
Its like you say, excitement not brewing for the 15 to 30K. But maybe an executive is going to have a nice IPA or pour himself some nice high-grade bourbon from an expensive decanter tonight because reducing costs may help the company achieve some metrics that get him/her a much nicer bonus.
Over the last 5 years Amazon has gone up a little more than a bond. AI maybe helps them change their situation, improve profitability, become a more attractive equity to own and buy.
There is a lot of talk about the impact of AI and the benefits and for the common man, the lower level employee, the non-executive, I think there are a lot of potential negatives. But just in the stock market, think there is a lot upside. Does not mean we'll be smiling when all is said and done. And I've seen Ex Machina enough to be freaked out a little by the potential. We may have a chance to make some money on equities but if we are not financially set up, the total net gain of AI could be weak. I am going to just try to make some money before they decide they can use some scripts to do most or all of my job. Also work hard enough to maybe be the last man standing if they only need one person.
@UlyssesExtravaganza said:
I am going to just try to make some money before they decide they can use some scripts to do most or all of my job. Also work hard enough to maybe be the last man standing if they only need one person.
That situation is akin to Nuclear War in that the lucky ones die instantly in the initial blast . Take this from someone who was in that position in a firm 22 years ago - in my case it was 3 of us left standing from a department that was a bit over 600 people just a month prior.
Glad to say that soul damaging experience of remaining won't happen to me again. I've positioned myself well enough in stable and liquid investments that I will be first on line to say give me the severance package (52 weeks plus accrued vacation days)
P.S. Liquid to me means a few mouse clicks and the money is in my bank account or will be the next business day.
It's the singer not the song - Peter Townshend (1972) Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
@UlyssesExtravaganza said:
I am going to just try to make some money before they decide they can use some scripts to do most or all of my job. Also work hard enough to maybe be the last man standing if they only need one person.
That situation is akin to Nuclear War in that the lucky ones die instantly in the initial blast . Take this from someone who was in that position in a firm 22 years ago - in my case it was 3 of us left standing from a department that was a bit over 600 people just a month prior.
Glad to say that soul damaging experience of remaining won't happen to me again. I've positioned myself well enough in stable and liquid investments that I will be first on line to say give me the severance package (52 weeks plus accrued vacation days)
P.S. Liquid to me means a few mouse clicks and the money is in my bank account or will be the next business day.
Can appreciate that angle. To spin it to cards, I was at a show talking to a dealer and the conversation go to she was selling the cards she was or trying to cause she was unemployed. Govt. Then the conversation got to other people we know who lost their govt jobs. Then the guy next to me says the worst thing is still having a job, in his case temporarily not being paid, doing all the work of the people who were not essential and could not come to work.
So yeah, maybe no last man standing wish. Or a very specific particular type of last man standing. Check to make sure AI did the work right. Hmmm. Let me see. Yup. Click. Then get another cup of coffee.
@craig44 said:
lots of doom and gloom out there...
Yes and deservedly so. Consider yourself blessed if you livelihood can not be simply taken away from you due to politics or corporate greed.
I do hope your comment was not glib or callow and that you possess empathy for so many who are less fortunate. St Peter is keeping your ledger even if you are not.
It's the singer not the song - Peter Townshend (1972) Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
Comments
It’s not about replacing people but anytime something can outperform a person it’s time for people to move aside and focus on another task.
We're having a hard time figuring out what these other tasks are. Maybe someone should design an AI to figure out what other tasks humans should be doing.
Most of the time when I get an AI on the phone, I'd rather be talking to a human.
At least one major US Bank has invested approx 20 Billion into AI/Machine Learning in the last 12 years , yep prior to it even being "fashionable" to do so. They would not spent that kind of coin if getting rid of employees was not at minimum their secondary objective.
It's the singer not the song - Peter Townshend (1972)
Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
Support bots are hardly AI. Most are script based and have nothing AI related.
You sound entrenched in your beliefs.
They are AI.
In 2023, this was the definition of AI that the White House used while drafting it's executive order on AI:
b) The term “artificial intelligence” or “AI” has the meaning set forth in 15 U.S.C. 9401(3): a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. Artificial intelligence systems use machine- and human-based inputs to perceive real and virtual environments; abstract such perceptions into models through analysis in an automated manner; and use model inference to formulate options for information or action.
Sometimes phone support bots decide to hang up on me.
I don't know if you remember Ask Jeeves? Search engine in the late 90s that allowed you to ask questions in plain English. Parsed English based on a language model to figure out what you were asking.
All of these AIs are following a script. It is a more complex script than a phone support bot's script. That's just how computers work. Stuff like the ChatGPT is using cashed websites and spitting out a formatted mashup.
None of these AIs are doing anything creative. They can write a song. Writing a song is just playing notes though. Anything can be a song. It's original as long as it doesn't match anything in the database. When you think about it, it makes being a human seem simple and stupid. Computers have been able to write songs since the 80s. Maybe even earlier.
I think most of the AI news out there is completely fabricated. If there is something amazing out there I haven't heard of it yet.
I’ll answer any real questions about AI > @1982FBWaxMemories said:
The stage coach driver union agrees with this message.
^ overly simplistic comment especially from the only one here who seemed to know the meaning of "japery" without having to look it up.
It's the singer not the song - Peter Townshend (1972)
Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
Well. It’s a long and nuanced discussion and I don’t have all the answers and this isn’t the environment. I wasn’t meaning to be trite.
Well AI is an umbrella term. At best chat bots are using RNNs. Fairly antiquated.
I dont have any problem with that description for that context. It makes as much sense as it needs to.
I’m sorry that happened to you. You didn’t deserve that.
Ask Jeeves was not using a language model as would be considered today. It was a keyword tokenizer.
So that’s not how LLMs work and you’re conflating that with the synopsis they provide based on what they are able to interpret.
You’re again focused on LLMs which are a very specific implementation of a neural network.
You’re speaking at a very high level on a wide ranging topic. There is a lot of BS out there but the only thing you’re touching on is LLMs for customer service. I’ll agree that’s boring.
You're attempting to make this sound a lot more complex than it actually is.
What am I trying to make sound too complex? We’re not discussing anything specific enough to be complex. It’s such a huge topic and we’re barely past “AI”. For 30 years the best we had was an encoder-decoder architecture and it was right to mock AI as merely programmed response however complex the pipelines got. Transformers and the introduction of context has unlocked NLPs to become smart enough to hang up on you… and perform many other complex tasks which people have struggled with. While automation gets the most news the real value, I think, comes from the ability of the convolutional layers to deduce the most effective correlations across remarkably large datasets without requiring brute force compute.
They're not being smart when they hang up. That's just what they were scripted, programmed, to do in that situation.
Computers are really good at going 1101000100100101010001001010111100101010101001 really really fast. That's really all they know how to do. Somebody tells them to do something when they a particular string.
Artificial Intelligence is an interesting term. Artificial, fake, Intelligence. It's a lot of smoke and mirrors.
They've been brute forcing for 40 or more years now. Some of these companies have brute forced the entire internet.
Humans having real intelligence isn’t a meaningful advantage. Our advantage is consciousness. Intelligence is a measurement of the rate of knowledge acquisition. In that sense, compared to many humans, machine learning is stupid as it lacks enough memory to maintain coherence. But we don’t need a general knowledge solution to make use of machine learning. And it’s much more than beep boop here’s a stream of bits. Don’t pretend you know what you’re talking about. Moral of the story is… don’t worry about buzzwords. If you’re hung up on how smart it is you’ve missed the point 10 stops back.
Don't pretend to know what I'm talking about? Bro. You're Mr. Buzzword in this conversation.
You should step back and think about what those buzzwords are describing.
Everything in computing is strings of bits. That's all computers do. High or low.
Don't you love it when someone uses the low (April 5) to gage returns when ydt is actually around half that? Good discussion though. My 2 cents.... data would indicate that equities are overvalued. Historically, after 3 years of double digit returns it gets a little dicey. Fed cutting rates bodes well for short term gains. Inflation kind of under control. Jobs market and consumer sentiment probably mystery drivers. Keep it simple and diversified. Time in always beats timing. Sport cards have zero bearing on market performance.
I'm convinced. Time to go all-in on NFTs!!!
I agree with this. It's easy to get carried away in forums but collector's items such as cards and artwork are always downstream of stocks and the broader economy. They are generally correlated with crypto as well (with a couple exceptions, I think of most cryptocurrencies as "collectibles"). This framework is helpful for understanding these market as reservoirs for when excess liquidity from the broader economy spills over.
The lone exception to this thought framework was Q3 2020 - Q1 2021. That was a one-time boost to this market that transcended historical patterns and brought new entrants back in. As someone who was selling unopened I'll never forget it. I started buying back in after National in summer 2023. The only way you get a true feel for the market is to participate. Generally speaking I felt I was much too early to buy throughout '23 and 2024. In crypto some call this a PVP (player vs. player) market. You are trading against established market participants and cannot count on new entrants as exit liquidity. Only in 2025 did the collectibles market get stronger which I would attribute to broader wealth effects that result from stock market at ATHs. Easy to identify with precious metals and crypto also at ATHs.
The main headwind is still lack of new entrants. Big questions - what few new entrants come in - where will they spend? And do scarce conditions exist in that market? I continue to believe there will not be a catalyst to bring new entrants to the hobby other than those who are "aging" back into it (most hobby entrants/returning story goes that they hit 40 and reached a level of wealth that allowed more discretionary spending). Saw good price growth in 1996-2000's unopened, MTG and Pokemon (late millennial products). That tracks with the demographic of those "aging" in to the market.
Despite the improving market, as an investor you still have the ongoing challenge of choosing the right product (player, or year/brand in unopened). Some of my unopened stuff has doubled while admittedly, some of my "investment" unopened from 2023 only kept pace with shipping price increases (inflation). Kind of sad, but it's easy to make mistakes in the hobby. And as I have stated before, you always have what I've often referred to as the 20% hurdle when liquidating. Generally this is 13-15% eBay fees, 3-5% shipping cost, 3-5% slippage (returns, theft, act of god). And when all is said and done, capital gains tax.
At the end of the day I still do this for fun/completing collecting goals, though there is an investment mindset behind purchases that cannot be ignored.
Binary is a representation of data but everything in computers is not high or low and that’s not all computers do. Binary states are foundational for classic computing but that’s like saying humans are just cells and nothing more. If you have a pile of cells well you’re looking at a human. That pile of leaves over there? That’s cells bro. Thats all humans are.
My area of expertise is in vision transformers so if you want to discuss details there that’s what I’m most interested in. It’s also most applicable to what PSA should be investing in.
There is definitely still a lot of confusion out there regarding AI. If you want a little insight into how a LLM works I highly recommend this 30 minute video (one part of a really good series). It's a little dated but I love it and have shared it with several family members who had thought this stuff was magic. It isn't.
https://youtu.be/wjZofJX0v4M?si=ON81NJQiRSvCrW1u
I think that is exactly it. Its not as if we are waiting for it to happen either. Its already happened. Ask a kid fresh put of college wanting to get a job as a Jr. Programmer how wonderful the experience of trying to find a job now is. What people are doing with AI right now is finding tasks/jobs that can be done with AI so they can let people go or not hire them. They can spend less not paying employees and their salaries and their fringe benefits and they can rent one less floor of a building. Then the AI can do the work in minutes that may have taken the person days weeks months. So for now saving money on labor and the end result is improving profits and beating the # when they report or just improving year over year on the numbers.
To clarify, I am not saying yay, this is a wonderful thing and lets celebrate its arrival and all the consequences. I do think it has some bad potential for the working individual. But just like a wager, you may despise the Chiefs but do you think they will win. If so, bet the Money Line.
And I think right now is just the tip of the iceberg in terms of seeing how much it can do. I just mention one thing where it can benefit companies in terms of decreasing costs while getting similar or better production. But we had the ChatGPT moment. Which happened with x amount of compute power and older less powerful Nvidia chips I would think. What is going to happen when all these data centers are built and powered? What additional progress will be made?
But I think what is interesting is between now and then so much money will be spent to build these data centers. They dont even have the energy to let all of these planned data centers run now so they are reopening shutdown nuclear plants. The spending impacts so much more than a company like Nvidia who makes the chips because you cant run a data center with an Nvidia chip. Cant just throw it a patch of dirt and say work. Think things. Answer questions. You have to build the building. Construction. Materials. You have to have the servers. You need the equipment, cables, storage. Cooling. Energy. It would be really cool to see the Accounts Payable list of vendors for Microsoft Amazon Meta Oracle Google.
If AI ends up being a bust, and I dont think it will, I dont think we would see it until 2029 2030. The buildout will be enormous and the nuclear power needed to fuel so much of it may not be ready until 2027 2028. Building a data center is like building an MGM Grand too. Cant get that up and running overnight. Until all the building is complete, which may be never or a ridiculous amount of time cause its like building the railroad or the highways, all the purchasing and building can be enough to make some companies really successful for 3 years or so. Then reevaluate and maybe its time to sell but it could have been a nice ride.
I think that is part of the issue too. I think a lot of the people who dont fully grasp it say, I dont get it. Its a bubble. But to really have that opinion that people want to listen to and trust, I think they should get it more and then say I fully understand it and I think its a bubble. I'm not saying people here. Say whatever you want. But more people who are analysts who invest in the financials and healthcare and consumer staples and that is their wheelhouse but then they have a strong opinion on tech. I think its popular to call it a bubble and its getting trendy.
I will admit some people knee deep in tech still say it. Like Sam Altman. But maybe what they feel the value of his company is is overinflated and therefore sure, why shouldn't he think its a bubble. If his company was public possibly the most bubbly stock out there.
I do listen to the guys who are knee deep in tech stocks who specialize in that area and understand it well and most are enthusiastic and dont feel its a bubble. I think its the other guys who are less involved, more diversified with less tech in the portfolio who say that its a bubble.
>
>
High and low is actually all computers do. You're typing a lot but you're not saying very much. You keep on saying you want to talk about it.
I don't know if anyone else has noticed this, but, ChatGPT is mostly doing the same thing your search engine does. Except, instead of serving you a link to a page with information on it, it's serving the data from the page, mashed up with other data from other pages on the same topic, and typing out that information really slowly, to make it seem like it's more human.
A few years back all these companies started complaining that these AIs were being trained using the companies data. What's happening is, companies like Google, who have been crawling the internet for 40 years or so, have cashed the whole net. They started off making a directory of web sites. But, then they started saving the sites and scraping the databases. For instance, a lot of the time you don't have to click the wiki link because the information from wiki shows up on the right panel of your search results because Google saved the wiki page. They're saying their AI learned it. As if their database is a brain. It's similar to me recording the world series with my DVR and saying my computer remembers watching it. Then distributing the world series and saying my AI is talking about the world series from memory.
I don't disagree. Analysts should be separating the wheat from the chaff. There are companies which are BS and circle-funding to drive valuations. There are companies out there who have no idea what they are doing. We're at the point now where there are tools out there which allow anyone to start labeling and start training so there's a lot of poor models out there that companies are building their valuations off of and these modules are destined for early plateau or complete collapse.
I think Altman has embraced his own confidence and he's started to make stuff up (chart-gate)... GPT-5 has regressed for sure. I don't know if it's reasoning collapse or just F1 divergence but they are already raising money... I think like $2T to build the next generation model. We observe that these LLMs tend to get over-fit with relative ease.
To the point about jobs, loss, shift, however it shakes out here, there, and everywhere... this technology doesn't do much unless there is economic activity to support, so there's that.
OK. let's do 1 thing at a time.
High and Low is all computers do.
What do computers do.
They take input. They store data. They process information. They produce output.
This is more accurately what computers do.
How do they do it gets us closer to binary and why it's used and so on and so forth.
That's good. I studied at the University of Minnesota Twin Cities, obtaining two bachelors degrees, one of which was in Computer Science, before attending graduate school at MIT where I also worked in MIT's Lincoln Labs. So sure... feathers out huh.
You're describing the concept of abstraction. The more you squint the less specific something looks so it's reasonable to describe LLMs and Keyword search as similar. They both take input... parse it... process it... and produce outputs. One is just fairly simple and the other produces better output. They work very differently under the hood.
What is this fixation on LLMs. We should all know that they are a NLP which produces a scraped synopsis which they value based on learned interactions. And I won't go past buzzwords here because I can't fit a talk on the concept of Perceptrons into a forum post.
It's not a technical statement to refer to a database as a brain. I can't comment on a statement I haven't see or reason why it was made. No one that matters thinks a database is a brain.
The data that companies like Google captured from all those years (20 or so) of user-interaction is priceless. It's labeled data and that's like a pile of money when you're training a model. But what's funny is that when you populate your layers with this data and the proper weights based on that pile of empirical data you start to get a little bit of reason out of your model and as you curate your training data-set you continue to increase the amount of reason. We look at our probability histogram to understand whether we're increasing or decreasing reason. It's much more than a keyword search and a rank-order output function.
My best advice is to question what you think you know. With your background and an open mind I suspect you'll be able to see it more clearly if you take the time to explore. I didn't consider AI much more than a statistical process to filter to the most likely answer for many years, and it wasn't much more than that in practice. We were telling it what to do in the convolutional layers. We were telling it what the important variables were. The big difference really came with the introduction of the transformer architecture.
What are you arguing here exactly?
You know, rank-order, that is weight.
I would have thought your advice would have been, "lie."
I suppose, if I was arguing, my position would be that there is more to machine learning that statistical process.
With keyword search those weights are static or, at best, periodically updated system-wide. We used to refer to these are hot spots and we would adjust to produce or reduce hot spots in the results.
Rank order is output weight. These weights which are the non-linear part of the data series propagation are input weights which themselves are further perturbed by the recursive redimensioning of the convolutional layers based on the feedforward networks. We encounter this process anytime someone says something that can be interpreted in multiple ways. We think about these possibilities and use the context we have to reform and inform. This is what FFNs and RCLs do. Instead of a keyword search that ranks output they go through each layer to an output and then pass that information back to each layer independently to reinforce learning and produce more accurate output.
Why would you have thought that?
Thanks for the input/education bgr and others. I will admit I have no hands-on programming tech experience. For my second job out of college I was fortunate to be able to get a Jr. Sales job in Software for a tiny company with about 11 employees. I didn't have a clue but loved the experience and challenge, pre big-internet, of absorbing as much info as possible to be able to have conversations with people and pitch the product.
Second job was more of a reseller of Oracle products. So had to understand just enough about databases to get by even though it was a little more transactional order-taking. But cool to be a part of any product overviews. Back then it was the database licenses being sold and then Developer 2000 Designer 2000 tools to help write applications because in 1997 or so 2000 had such a cool ring to it. Had some other jobs after that in the tech sales space tech training space. Then there was 9/11 and budgets froze and selling was tough and in time I realized I was an average but not outstanding salesman and time to move on.
But my interest in tech companies has remained. I dont have a diversified porfolio. Or if I do, it only looks that way with energy and construction etc. but you still can tie it back to tech with the data center connection. Portfolio is close to 100% tech and has been since about 2018.
So my angle with all of this is mostly business case. I am not able to get down in the weeds. Just looking at the business case and can see that a lot of the companies with the largest market caps believe in AI and cant spend enough which I think for at least a while that means a decent # of companies win. Certainly there are a number of frauds. Just because you say AI 10 times in your earnings call wont make you a winner. Just like Pets.com. Having an internet angle wont make you a winner. Having an AI angle wont make you a winner. So as always, pick right. And there is risk. People dont have to play if they dont want. Folks can stick to value investing. But on the original point just because tech stocks aren't like value stocks, I dont think that should have value investors calling tech a bubble. They are just two different worlds with different rules.
What gives me pause is how much the overall economy appears to be relying on the rampant investment in artificial intelligence. When I look at the model failure rate of LLMs of 80% and general models around 95% that gives me a bit of pause as there's a lot of ROI not being met. A big part of this, from where I sit, is that there's not enough people who know what they're doing to cultivate these models effectively. This is the 'data scientist' role which can only be ill-defined. Maybe it would have been called a statistician more generally when I was coming up - someone who can put data to use using maths. To the point CardGeek was making about being able to make sense of the data for useful purpose - there is a lot of companies flailing around doing nothing useful. There are always frauds... just like dotcom.
China is doing what China does and putting data centers underwater. Microsoft actually tested this - Project Natick, they determined it was not feasible - many issues - obvious ones if you've seen any movies where people build stuff underwater I guess. There's a lot of cool stuff going on in data-center development by companies like Microsoft, Apple, IBM, Oracle... I think that's a really good area to watch for the pulse of how things are going. I also think that this is going to push development in our energy infrastructure here. If everyone can find a way to go-along and get-along here there's a bright future. Outside of the larger geopolitical brush strokes that is where I would train my eye to see where the success needle is for the "AI Economy".
When I saw 40 years I had to count my rings... I remember talking about BackRub on EFnet #linux... Google was BackRub before it was Googol before it was Google.
One of the most transformational changes to search-algorithm that Google invented was to follow the links with PageRank. Such a simple thing but it had such a major impact on the quality of result they could provide from their search that they won the market.
But I would not say that it is true that Google scrapes databases. They do save sites in Google Index, including raw content that is accessible. I have found that Google's spider does respect robots.txt for Search Result, and noindex for Indexing if done correctly, but this is not the case everywhere and has not always been the case.
man, a lot of techno talk for a card forum...
George Brett, Roger Clemens and Tommy Brady.
Maaa baaad
Sometimes ya gotta brew your own excitement when the other brewers don’t.
Agreed it is a long and nuanced discussion BUT the exact example I referenced eliminating jobs was an objective for that firm! They tried to hide their glee but IMHO it's still present
https://finance.yahoo.com/news/jamie-dimon-declares-jpmorgan-chases-173130430.html
AI's Impact On Jobs And Roles
"People shouldn't put their head in the sand. It is going to affect jobs," Dimon told Bloomberg, warning that AI adoption will reshape the workforce. He said automation is already streamlining operations across the bank and added that "there'll be fewer jobs in certain functions," while emphasizing the importance of retraining and redeployment.
CFO Jeremy Barnum said during the bank's annual investor presentation in May that the firm is "asking people to resist headcount growth where possible" while still hiring in "high-certainty areas."
BTW this discussion is a lot more interesting than the self-serving attempt to pump cards as the OP is trying to do! @bgr I'd like to think we are total agreement about that.
It's the singer not the song - Peter Townshend (1972)
Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
Have any of you ever read Hitchhiker's Guide to the Galaxy? It's the funniest book I've ever read. The book has been adapted into a Hollywood movie and a BBC TV miniseries. While the BBC series is a better representation of the story, neither of the two do the book justice.
Interestingly, the movie cost over 100 million dollars to make. The BBC series probably cost about 50 bucks and 2 large pizzas to make.
In the book a huge computer is built to calculate the answer to the ultimate question of life, the universe, and everything. The AI tells the scientists that the problem will take 7.5 million years to process and to check back later. Scientists maintain the system for 7.5 million years. After that time the AI tells them that they're not going to like the answer and asks if they really want to hear it. The scientists insist that they want to hear the answer. The AI tells the scientists that the answer is "42".
42, for two, 4 + 2 = 6, six, sex.
You kind of have to read in to it, if you know what I mean.
Sometimes people write a lot that's not really worth reading. Dougles Adams is worth reading.
Excitement is not brewing at Amazon, at least for the 15-30k corporate employees that are being eliminated.
Tip of the AIceberg
I personally love Hitchhikers Guide and, sadly, I've quoted Douglas Adams on these forums many times without recognition.
I get that technology takes jobs is a bad thing. I'm on your side. I supported the Farrier's Union! Down with the Automobile!
I also fully support PSA replacing all those terrible graders with machine eyes!
I bought a Gold Nugget in 2024. It glows.
Here's from the president of Upper Deck on LinkedIn. Maybe a bit of sour apples because, other than hockey and oddball titles, not sure how relevant they are in the market now, but I think this is a fair take:
Everything speculative is up right now: The stock market. Bitcoin. Trading cards…and it's pricing some collectors out.
Here's what the data shows: 50% of consumer spending in the US is coming from 10% of the population. Rich people are carrying the market. Not just in cards, but across every category.
Concert tickets? $1,000 for nosebleeds. Luxury goods? Selling faster than ever. High-end collectibles? Moving at record prices.
When boxes cost $1,000 or more, you're not buying for fun anymore. You're speculating. You're betting that what you buy today will be worth significantly more tomorrow.
And that works. Until it doesn't.
Sports cards have always had speculation built in. You collect a rookie hoping they become a Hall of Famer. But what's happening now is different in scale. And when markets become this speculative, they become fragile.
I'm not saying the sky is falling. I'm saying we need to be honest about what's happening. This isn't sustainable growth driven by collectors who love the hobby. This is wealth concentration driving prices that most people can't participate in.
And if that money suddenly moves elsewhere to bet on something else? We all will feel it.
Gold nuggets are up too.
That is where that steaming hot cup of Excitement Brew comes from. > @80sOPC said:
True. And I heard some discussion this morning, not for the first time, that its not just companies laying off but deciding they dont want to increase head count.
Its like you say, excitement not brewing for the 15 to 30K. But maybe an executive is going to have a nice IPA or pour himself some nice high-grade bourbon from an expensive decanter tonight because reducing costs may help the company achieve some metrics that get him/her a much nicer bonus.
Over the last 5 years Amazon has gone up a little more than a bond. AI maybe helps them change their situation, improve profitability, become a more attractive equity to own and buy.
There is a lot of talk about the impact of AI and the benefits and for the common man, the lower level employee, the non-executive, I think there are a lot of potential negatives. But just in the stock market, think there is a lot upside. Does not mean we'll be smiling when all is said and done. And I've seen Ex Machina enough to be freaked out a little by the potential. We may have a chance to make some money on equities but if we are not financially set up, the total net gain of AI could be weak. I am going to just try to make some money before they decide they can use some scripts to do most or all of my job. Also work hard enough to maybe be the last man standing if they only need one person.
That situation is akin to Nuclear War in that the lucky ones die instantly in the initial blast . Take this from someone who was in that position in a firm 22 years ago - in my case it was 3 of us left standing from a department that was a bit over 600 people just a month prior.
Glad to say that soul damaging experience of remaining won't happen to me again. I've positioned myself well enough in stable and liquid investments that I will be first on line to say give me the severance package (52 weeks plus accrued vacation days)
P.S. Liquid to me means a few mouse clicks and the money is in my bank account or will be the next business day.
It's the singer not the song - Peter Townshend (1972)
Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025
UE - correct, some very large companies are in hiring freeze. PWC is apparently going to cancel 100k open headcount.
And to the post from UD’s president, luxury goods are not on fire. Look at luxury stocks, most are down for the year.
Something is brewing and its not excitement.
“This is not, however, an ideal Universe”
I'm very seriously considering an offer on about 65% of my collection. Feels like good timing and I'd like to try the money somewhere else.
Can appreciate that angle. To spin it to cards, I was at a show talking to a dealer and the conversation go to she was selling the cards she was or trying to cause she was unemployed. Govt. Then the conversation got to other people we know who lost their govt jobs. Then the guy next to me says the worst thing is still having a job, in his case temporarily not being paid, doing all the work of the people who were not essential and could not come to work.
So yeah, maybe no last man standing wish. Or a very specific particular type of last man standing. Check to make sure AI did the work right. Hmmm. Let me see. Yup. Click. Then get another cup of coffee.
They should probably replace the IT guy too.
Will the last employee leaving kindly turn the lights out.
lots of doom and gloom out there...
George Brett, Roger Clemens and Tommy Brady.
Yes and deservedly so. Consider yourself blessed if you livelihood can not be simply taken away from you due to politics or corporate greed.
I do hope your comment was not glib or callow and that you possess empathy for so many who are less fortunate. St Peter is keeping your ledger even if you are not.
It's the singer not the song - Peter Townshend (1972)
Not even a minute do I buy the whole buh buh buh I'm a man-child japery - Me 2025