This is what we call a solution to a problem that does not exist.
2 Likes
I saw you mention it - but are they still doing that in any material way? All the stuff I could find on that was back from the âPeak Blockchainâ era of everyone trying it out so as not to be left behind. Plenty of blockchain enthusiasts pointing to WMT, but nothing in the last few years to indicate theyâve given it any real effort.
The blockchain is useful for tokens and alt-currencies. No doubt. But almost all the other uses - the ones that were going to fundamentally change the world - ended up being abandoned as most companies wound down their pilot programs.
2 Likes
Just because you donât know⌠the entire US supply chain is adopting blockchain, many companies donât have qualified people to do the implementation.
1 Like
Let us break it down⌠when a new technology comes, there will be lot of pilot programs to learn the technology, and find better fit. The very nature of pilot program is a small percentage only will go to production. I am not sure which âchange the worldâ you are talking aboutâŚ
I recently talked about how US converts everything into a financial asset, securitization, then it is going to be tokenized. Technology adoption has its own phase.
Folks here make lots of statements on hunch, and based on their belief, not with some real data behind in it. For ex: Blockchain is implemented and in use in 80 of the top 100 US companies. That should tell you something about the adoption of the technology.
1 Like
Iâm not really sure this is true. In fact I doubt this with all the fiber of my being!! But then, Iâm old and antiquated and would just like everyone to get off my lawn.
JimA
3 Likes
Okay - so I decided to google that statement (as I find it ridiculous) and I see it is attributed to Cryptonary (clearly no vested bias) - and the AI summary said this: The statement that blockchain is implemented and in use in 80 of the top 100 US companies is misleading and likely refers to outdated or misconstrued data.
JimA
1 Like
It does - it means that everyone dabbled in it. It doesnât mean the technology is having any kind of material effect on their operations. For example, I would imagine that an overwhelming number of the top 100 U.S. companies have some sort of onsite âgreenâ energy program - solar panels on some of their buildings, maybe some Green Roofs here and there. But very few of them are getting any material amount of their energy from distributed renewable technologies like that: theyâre still going to be getting nearly all of their power from the grid.
I have no doubt that lots of companies still have some blockchain stuff that they tried and didnât get rid of, but the âPeak Blockchainâ era came and went with very little implementation of the technology in nearly all of the areas that it was tried.
Most blockchain projects never delivered real value. Companies rushed in, driven by fear of missing out and the promise of technological transformation.
But the tech wasnât ready, and the solutions it supposedly offered were often misaligned with real industry problems. Companies tried everything, from tracking pet food ingredients on blockchain, to launching loyalty programs with crypto tokens, often without clear benefits or better alternatives.
In the end, about 90% of enterprise blockchain solutions failed by mid-2019.
The AI hype is just like the blockchain frenzy â hereâs what happens when the hype dies
4 Likes
Way back when I worked in NYC, to help an organization migrate from AS/400 to Oracle Database, I told folks I need to take the server/ database down for couple of hours to apply a patch and also do an OS patch/ upgrade. The team was horrified, and the lead AS/ 400 programmer demanded to know,
âwhat kind of garbage system you are dumping on us, how can you apply OS, database patch and restart a server in 2 hours??, it will take us anywhere between 8 to 16 hours to restart AS/ 400, how can you do this in 2 hours?â
Now, at that time he was a very experienced, and very well respected guy within the organization. But, he had very little idea of Unix or Oracle Database, and couldnât understand why we are upgrading to newer version within couple of months of starting the projectâŚ
When something is new, it quite natural for people who are very familiar with old ways of things to look at it with suspicion and question why we need it. Human natureâŚ
Separately, like a broken record, companies like Wal-Mart are adopting technology at a high degree⌠we are not aware of it⌠outside of this news, $WMT already has sensors in every refrigerated session to check the temperature, and when they sense a malfunction, will automatically dial the HVAC team⌠not the store manager.
it provides a new stream of data into AI systems, enabling them to be even more effective in giving Walmart greater visibility into supply chain operations.
The technology initiative is already making a significant impact by eliminating some manual tasks and providing automated alerts, Cathey said. âAssociates no longer need to perform time-consuming checks to locate items,â
Having a root-top solar panel may satisfy a household electricity needs, but not necessarily, an office building, factory, or even a mall. But that doesnât mean they are not implementing it. Now, most retail REITâs have root-top solar panel as a standard feature.
It is great to have these conversations, but it looks like we are having this conversation while living in two different planets 
1 Like
Nobody is denying that companies are always adopting new technology, sometimes at a very high degree. That simply doesnât mean that every new technology will get adopted at a very high degree. It doesnât even mean that every new technology that people are really very hyped about will get adopted at a very high degree. The syllogism:
- Companies adopt a lot of new technologies.
- X is a new technology.
- Therefore companies will adopt X.
âŚdoesnât hold. Companies adopt new technologies if those technologies are useful for them, and not all technologies end up being useful for as many things as their advocates claim or hope.
AI - which at this point is almost being used as a synonym for âsoftwareâ - will end up being used all over the place. Because again, virtually everything can be labeled as AI.
The narrower question is whether the scope of optimism is justified. Sam Altman supposedly plans for $5-7 trillion worth of infrastructure to build and run whatever AGI agent he thinks he can build - which is much more money than exists in the entirety of venture capital. For that to actually happen - and for OpenAI to live up to those promises to buy trillions of dollars in chips and cloud computing that are inflating AI company valuations - they will need to have AI applications that make a ton of money. Like, more money than one can possibly imagine kind of money. So itâs not a bad question to ask whether what AI can do is worth a whole lot, because if itâs not then people (even companies) arenât going to be willing to pay a whole lot for it.
6 Likes
Easy: âthis time is differentâ. Again. 
3 Likes
This is about switching/adoption rates in general industry. For investments, this means following the value proposition - and business results.
Upthread we see:
âSolution in search of a problemâ
âIf you build it, they will comeâ
âBenefits with no direct tie to financial valueâ
and several equivalent statements.
Do these statements make investments profitable? It dependsâŚ
Itâs case specific, so each company should be evaluated on itâs own merits.
We should just rename this thread to: âI want to make generic investments for the AI space. Itâs so new, I am finding lots of conflicting information and cannot make a decisionâ
Salesforce (CRM) has been a laggard this year, losing more than 20% at times to peers. Is this due to marketing (hype generation), strategy or approach? They have their annual conference in San Francisco where Mark B. is revving up the hype machine to signal greater AI involvement.
Many companies are doing this same thing (lagging behind the Mag 5-8 mega caps), but on the journey to value none-the-less.
The results for each company should be evaluated to see âif they followâ.
1 Like
The problem thatâs being addressed is similar to the problem the Manhattan Project addressed. Letâs call it the âHoly crap, this thing is powerful and can be weaponized! Weâd better develop it before the rest of the world figures it outâ problem.
Iâve got a feeling that Chinaâs advancements in AI are not being leveraged for chatbots and virtual girlfriends. Weâre so screwedâŚ
5 Likes
Yes! For investment purposes itâs finding the companies that can monetize AI, monetize above and beyond the huge cost of AI.
What caused the dot-com crash was not being able to monetize the web. One of the Web cheer leaders published a paper to the effect that profits did not matter as long as you could get venture capital cash. The paper was quite compelling. I showed it to my stock broker, an MIT alumnus, and he just shook his head. I have tried to retrieve this paper from the web but it seems to have been pulled in shame.
AI is third generation software. There are two ways to sell software, as a stand alone product or as the heart and brains of hardware or service. The iPhone and Google search are probably the best examples.
The Captain
2 Likes
Or rather, not being able to monetize it enough. Companies made money on the web. They just didnât make enough money to justify the money they were spending. The VC investors were fronting the money that they were spending on the web, but eventually they had to start earning enough money to justify the investments. They didnât, and it was eventually shown that they couldnât, so the VC money stopped - and then the music stopped.
In the current AI boom, OpenAI has basically agreed with a bunch of companies to spend about $1.2 trillion dollars on infrastructure (chips and cloud services and whatnot) over the next few years. Thatâs a huge amount of money. Thatâs probably more than all the venture capital money in the entire world over that time frame (itâs about $370 billion per year globally). The deals where OpenAI promises to spend hundreds of billions of dollars on Broadcom and Oracle and Nvidia stuff has driven those stock prices to the moon - but at the end of the day, someone has to actually pay for all that stuff. OpenAI doesnât have $1.2 trillion today, so it will have to come from somewhere.
Where? OpenAI is doing some fascinating financial engineering alongside their computer engineering. Matt Levine explains it better than I, so Iâll just link - theyâre fun reads:
5 Likes
Dead is dead by any other name.
Willy Shakes
1 Like
Is AI really new any longer? After all your Unix OS example was installed in two hours during one day and then was old hat. Letâs use your example after all.
I am trying to keep them separate!!
Agreed. Just anticipating the response that people are making some revenue selling AI products today. Thatâs certainly true. The question is whether whatever AI agent(s) come out of these multi-trillion dollar investments can earn the many trillions of dollars in revenue necessary to support those investments. Basically, AI will need to be several times larger than the entire existing global software as a service market is today for this to work. That seemsâŚ. optimistic, considering what AI is showing itself to be able to do so far.
6 Likes
Investing is not about technology, itâs about money. Even if AI is successful, at this time I think it will be successful, 80% of adopters will not profit from AI. Traditional investing strategies are needed, and by this I donât mean adoring P/E ratios. Much more relevant are things like market size, total addressable markets (TAM) and business strategy. What do investors think about selling rejects to poor people?
40 years of 21% CAGR, Ross Stores, Inc. selling rejects to poor people. BTW, the selling is not the major factor, itâs buying the rejects! i had a supermarket client and I could not figure out how them made over 30% return on capital by selling razor thin margin products. The head of purchasing told me their secret, âI have to sell the product three times before I pay for it.â In other words, if they get 30 day terms the item has to be off the shelves in ten days or less. This is not how Ross Stores does it, they buy bulk which they dole out in accordance with the seasons. At Dell they forced suppliers have their warehouses right next to the Dell factories. One advantage of software based business is Increasing Returns as opposed to Declining Returns of traditional business.
Donât worry about AI, worry about the company you want to invest in. You hate the CEO? Tough!
The Captain
Decades ago I hired a Sun Microsystem based web hosting service. When one day the equipment broke down they had to order parts special delivery from California which resulted in several days service outage. Lesson? Use only Intel based servers.
1 Like
Itâs about technology if the value of the company youâre investing in (directly or through the S&P 500) is a technology company whose value depends on whether a given piece of technology works or not. If youâre buying a retailer like Ross or Dell, then what matters is the margins on their resale of products supplied by someone else. If youâre buying a drug company, what matters is how good their drugs are. And if youâre buying a company that makes software or chips or technology products, the value of the company depends on the market for their software or chips or technology products.
The âAI riskâ in this thread is questioning whether the software products that companies like OpenAI and others are going to be modestly valuable or world-changingly âWeâre Building Godâ kind of valuable. What will AI be able to do by itself in one, three, five, or ten years? If the answer is, ânearly everything a moderately smart human can doâ then AI might be worth spending $5-7 trillion to build and run. If the answer is, ânot much of anything,â then obviously thatâs not true. And if the answer is âsomewhere in-between, but more towards the latter than the former,â itâs still going to be a major issue.
IDK. Again, total VC capital for all global investment everywhere for everything is about $240 billion. Not just all tech VC investment - everything. Open AI wants to spend $5-7 trillion to build and operate their AI Thing. Can AI Thing ever be worth spending $5-7 trillion to build and operate? Itâll have to be really good at doing stuff without making mistakes or needing supervisionâŚ.
8 Likes