There is the underlying reality and bias. When I'd pick up my kid from Kindergarten it was me and 20-30 other moms hanging out on the playground afterwards. So yes, in that sense we may not see gender neutral outcomes.
The bias however is absurdly strong. My first lawyer said I should have taken my ex's offer of slightly less than 50/50. She said that my ex's floor was 50/50 and that if we couldn't show that we could effectively coparent they'd give it to the Mom.
That's exactly what happened and it held through her completely cutting me off from my son. I even went through a modification try and the same Judge upheld the situation.
As a result I haven't spoken with my son in almost 2 years now.
I would definitely have preferred to have had a couple more days a week off with my kids while they grew up, and given their Mum a chance to do her thing, too. Instead, it became my responsibility to pay the bills, and the rest of the crews right to make them. So yeah, that's what I did.
>2 years
Kids need their Dads, and Dads have to be strong as hell for them, even still, in this modern world. I am super glad to have gotten my two giant oafs into adulthood well enough to be surfing the wonders the world has in store for them both .. but yeah, you'll get your ass kicked. Be a kick ass Dad, then.
The most important thing is that your son knows you're there for him as he grows up. Maybe keep a journal, don't stop writing important things to him, notes and ideas from his Dad, for him to read when he gets his own agency over things.
How is it even possible for someone to prevent you from seeing your own child for two years? That would be completely illegal in my country. Can one parent really just decide to cut the other parent out of their child's life and get away with it? It's absolutely proven than this is against child's children best interest.
How is it possible? Well, theoretically it's not. The order I got and subsequent denial of modification blatantly violate state law and appealate precedent. But it's all procedural. The standard for the underlying opinion of the judge is that if there is one fact to support that opinion then the ruling is upheld.
So I could have appealed and won. But then I go back to the same judge that doesn't think my kid should have any contact with me and all she has to do is find one fact to support that.
More often a custody order is just a piece of paper. The primary parent can screw with visitation easily and it takes time and money to go to court. The original divorce cost me $100k, another contempt motion $7k, and the failed modification $20k.
Custody ends for Fathers when they give up. I gave up unusually early (which royally pissed off the judge) and this is my punishment.
Family court is a scam, and if you end up there you’ve already lost. For some reason your usual legal rights don’t apply; the entire system seems to be constructed to drain everyone of money. It’s basically the definition of “the process is the punishment.” I’ve never seen a good outcome from it.
AI has made my job a lot easier and it feels like the fleeting interlude where expectations have not yet risen to match.
Actually using AI I think is a net positive. It frees up the drudgery and let's me think high level. I can try different ideas and throw out ones that don't work without expending much effort.
There's some loss with not being as hands on. But I feel like my impact is higher so I like the trade.
This is quite misleading. It includes numbers for commitments for leases of data centers and payments for power that are decades into the future.
Even if AI completely busts and all of these servers are worthless that doesn't mean the data centers are. There's still a need for data centers to host non-AI servers. The hyperscalers will close/sell other data centers and serve from places where they have long term commitments at.
3.2 trillion sounds like a lot until you spread it over 20 years and half a dozen trillion+ dollar companies. It's hard to tell how much a bag they're holding but it's almost certainly negligible compared to revenue and market cap. The only real systemic risk is the risk of for some reason not needing any data centers.
As I've been pointing out a little bit recently -- the 'downside' scenario -- somehow we conclude we have enough inference at our world's desired quality and therefore demand stabilizes below that of these datacenters -- would be GREAT for the world. Super cheap intelligence for all.
If some lenders go bankrupt on a datacenter note post-buildout and the capital structure is reworked, lowering the costs by removing capitalization costs from the cashflow of the inference provider, that is all to the good of people who did not do the initial lending.
I think that is extremely unlikely, but it's good for datacenter doomers to think through these scenarios IMO.
Your take seems a bit optimistic if we use the dot-com boom as an analogy. The dot-com boom sped up a massive infrastructure build-out, but also left damage that lasted for a very long time, with many organizations collapsing, and others taking a very long time to recover (e.g. cisco networks). It also seems that these 'data-centers' have a higher depreciation rate than 'dark-fiber' did post-dot-com, which means that there is likely to be greater capital waste.
I am not sure whether the dot-com boom was a net-positive, and the AI boom may be less productive.
It absolutely was a net positive. On the infrastructure side we got datacenters built and tons of dark fiber put down. At the start of the boom we had only 2G cellular for goodness sake. I think you could argue we wouldn't have necessarily had the rise in cloud compute without some of the 'overbuilding' of datacenters in web 1.
For reference this 'massive build-out' resulted in, according to a Berkeley study I found, 500MW of total datacenter power in the US across 320 locations at end of 2020. This wasn't too much for very long: as of 2005, according to Tier1 there were 511 datacenters. And it was likely only really a year or two: Dupont's VA3 utilization numbers published it at 50% utilization in 2003, 89.4% in 2004, then full thereafter.
Meanwhile, these infrastructure improvements allowed, for example, things like the iPhone to be built and deployable -- there was 3g, there were a LOT more web devs that could create the early (web only) apps the iPhone needed to compete. This in turn allowed another round of investment to be built on top of all of that.
On the capital side, investors were also able to buy Amazon at around $1 / share. AAPL at iPhone launch was $7/share. I haven't split adjusted either of these, this is just from memory.
I don't think a bust was necessarily good but the boom undeniably built out infrastructure, a small fraction of what the world is currently using, and did it pretty quickly, and then post bust that infra got bought out in a capital efficient manner. I think that's the bear case here.
Put another way, the Dotcom one boom was wrong about two things: timing was too fast, and the predicted scale of change was way, way, way too small. They undersold the future.
My point is that these hyperscalers need X traditional data centers and Y AI data centers. If Y goes to 0 that doesn't mean they eat everything they've spent on those data centers. They get converted to serve non-AI loads while other older ones get abandoned or leases not renewed and so on.
In that scenario the GPUs are a loss. But really at this point inference is so profitable it's hard to imagine these data centers not being able to at least cover operations & rent.
Ai datacenters are quite different from traditional datacenters. For example, a massive cost for ai DCs is power and the grid to support it. Regular DCs don’t need anything like the same power.
Cisco networks took a very long time to recover, but in the meantime, everyone else enjoyed massive growth in cheap connectivity.
Capital waste is only capital waste if the physical goods that were built are no longer useful for anyone. Otherwise, it's only a waste for the people who lent money.
This is something people miss when comparing things to dot com. It was a very thin layer on top of something new. 90% of the companies were speculative jokes with no basis in reality. Anthropic makes a profit. Google is a mature company that makes a profit and will continue to do so.
There's a lot of insanely silly AI products and services that have absolutely no use in the real world whatsoever. In that sense it's very very similar to the dotcom crash. And the crypto "blockchain" boom. And the metaverse. Remember when we all put blockchain on our CVs somewhere or we'd be unhireable? Who's talking about that now?
The Google and antrhropics of today are the ISPs of back then. They didn't go anywhere. It was more the services on top of that that crumbled because they had nothing real to offer. I see the same here.
In 2001 society wasn't ready. People didn't yet want to order their groceries online or read books off a screen. The computers weren't ready, they were slow, heavy and clunky. The smartphone still needed to be invented. Mobile networks needed to come around, the miniaturisation of GPS, cameras. These things can't be rushed and that realisation popped the bubble. In the end they did happen, but in their own time.
I see so much of that now. We must have AI in our product! How, why, what value does it add? Who cares! It must be AI or we're out of the race!! That's really the hallmark of a bubble.
Here again, the usecases aren't fleshed out yet (except coding). The tech isn't finished, it's all very hit and miss, hallucination, loss of control (agents hacking) etc. The people certainly aren't ready, we need time to adapt society to new technology and a couple years isn't going to cut it. But these investors aren't investing this much to see a return in 2035. They want it much sooner. Sooner or later that will dawn on them and they start pulling out.
You are correct of course. I was conflating their operating cash flow stuff. They are getting real money from real customers is what I should have said. Other much more knowledgeable people have broken down all the accounting in a way that I can roughly follow. I suppose my point is that people comparing them to pets.com et al are torturing the analogy.
Not sure when you last were in SF but there is probably half a billion in Series A capital within 1 mile radius of me raised by very nebulous AI start ups.
The downside scenario isn't AI equity getting zeroed out and lenders getting crammed down. It's the federal government deciding that AI capabilities are of national strategic importance and throwing trillions of taxpayer dollars onto the bonfire.
It really depends on the "somehow". The most realistic downside scenario, it seems to me, is that most people don't care too much about super cheap intelligence for all because there ends up being another bottleneck for most of the things we value.
Like, trains are an incredible technology which transformed the world, but very few people go around counting their lucky stars that they live in a world that permits such wonders. It's a train, it gets things where they need to go, and by the way isn't it so annoying that sometimes it doesn't quite arrive when the schedule said it would?
> The only real systemic risk is the risk of for some reason not needing any data centers.
How is this not a systemic risk? Asserting that:
> There's still a need for data centers to host non-AI servers.
Isn't true because a NVL72 rack consumes 23x the power of a traditional server rack which is why there's new construction. You could retrofit such a datacenter with normal servers but then you're taking a loss on the GPUs which is the majority of the cost.
It's not a systemic risk because outside of society collapsing we are going to use lots of data centers regardless of what happens with AI.
At this point the question isn't really if we're going to use the GPUs they're building out. The question is how much profit they are going to make. I've seen people show margins as high as 80% on inference. That's a lot of headroom before we see data centers being shut down.
It is a big assumption that society is going to need more data centers no matter what. That requires a gain in efficiency by using more compute power. Most gains in efficiency in the past have been from using fewer resources. Spending more or producing more doesn't make your market larger. Before AI, the future of data centers was low power, efficient cooling, cheaper costs. How many low power ARM boards you could stuff in a rack rather than how many GPUs.
As of writing, the most dense server CPU is the AMD EPYC 9965, which has 192 cores. It is possible to build a 2x CPU 1U server, so that is 384 cores per 1U. I am unsure what is the limiting factor: heat or space on the die (probably lots is just L1/2/3 cache). When you look at the interior of a vanilla 1U server, I don't think there is any remaining space to add more sockets (CPUs). Final note: SuperMicro also offers an Ampere ARM chip with 192 cores.
It is, by definition, NOT debt. With debt, you are given a lump sum of money. These are spending contracts, but any contract has the ability to be gotten out of, unless your lawyers are very stupid.
They mean you're obligated to the money. You can't back out. You're in debt to them. But, some of this does not include obligation, so fundamentally isn't the same as debt.
debt means you owe money. There is no requirement you get anything in return for taking on that debt. The definition, since you wont look it up, is "Debt is an amount of money, goods, or services that one party (the debtor) owes to another party"
If I commit to buying some future goods, and sign that I will hand over that money, that's debt. It's money I legally owe someone else. Counting that as debt is totally fair, not counting it is shady as fuck.
If I rent a house, I receive no money, in fact I receive no tangible goods at all really. Just the right to occupy a space under some pre-agreed rules. Yet, I now have debt.
Hmm yeah but those data centres will be full of equipment for an NPU heavy workload. That doesn't do much for other workloads. And the hardware is the expensive part.
It's a tale of two AIs for Google. Cloud doubled revenue and tripled profit, probably based off serving models. So good for Cloud.
Search increased revenue and profit which is good. Means that LLMs are not yet stealing revenue. But also bad because the massive spend on infrastructure doesn't look like it's going to come back as increased revenue.
People on the internet really overrate google's TPU advantage. OAI and Anthropic and others have been using TPUs as well for a long time now too. And they've also hired lots and lots of talent from the rest of FAANG, hyperscale compute talent and knowledge is also well distributed through all the top AI companies at this point.
Think they're going to be like Oracle. Databases were a valuable thing when they were new but became free with postgres. Like Oracle Anthropic et al will move up the value chain.
Switching cost for a DB and switching cost for an inference API are very different things. They’re going to have to create lock-in with something other than model intelligence if they want to play Oracle’s games.
Vendor lock-in is inevitable when negotiating enterprise contracts. Companies will sign one or two-year deals based on headcount, convinced they're going to save a ton of money. And a big part of the appeal will probably come from intelligence capabilities being bundled with services they're already paying for (looking at you Microsoft)
Its nothing like Oracle - for long time there was Oracle and nothing so the army of mostly politically influenced salesman/lobbyist who convinced local councilman its better to spend $250,000 on Oracle license, continue to this days because migration also comes with risk and cost. Here you have Chinese models popping out less than 24 hours later, and they are of similar quality at fraction of cost. Its totally different world that what it was in 80s.
Should be obvious why any correspondence related to something over an 8 year span is expensive. To fulfill that you have to read or at least glance at every single email.
You have to request specific keyword searches and go from there.
The publication isn’t being dishonest - they openly published (in a post linked from the source article https://doublesidedmedia.com/2026/09/30/editorial-the-eugene...) the text of their two requests including the date ranges. They just don’t seem to understand the complexities of doing a detailed record search with broad criteria.
Perhaps you're misunderstanding the story here? The search did find them easily. The problem, the police claim, is that the search produced 2,689,000 records. Before being published, each record has to be reviewed, as it might have sensitive information about people other than Mr. Peckels that should not be released. (They asked for all "communications, correspondence, and logs", to be clear, so many of these records are going to be random freetext emails rather than structured reports which might be more easily analyzed.)
At least in Atlanta it seems like a lot of production bounced to even cheaper locales as tax subsidies became a bit less generous. We're at like half our 2022 peak dollar wise.
The bias however is absurdly strong. My first lawyer said I should have taken my ex's offer of slightly less than 50/50. She said that my ex's floor was 50/50 and that if we couldn't show that we could effectively coparent they'd give it to the Mom.
That's exactly what happened and it held through her completely cutting me off from my son. I even went through a modification try and the same Judge upheld the situation.
As a result I haven't spoken with my son in almost 2 years now.
reply