Google and Microsoft will obviously remain. I have a hard time envisioning that OpenAI or Anthropic will go under - especially Anthropic, who are reportedly raking in billions from Claude subscriptions.
Just from my armchair predictions, it's not really any of the juggernauts who have to worry, but rather the many companies springing up to try SaaS offerings with LLMs at the core. A bubble pop there could certainly cause some strife, but I'm just not seeing the mechanism by which these too-big-to-fail tech companies and the heavily invested "frontier AI companies" are going to suddenly cease to exist.
I think the dotcom bubble is a fairly apt metaphor in the key sense that the web didn't go anywhere - just a lot of small players lost their tickets on the gravy train. "Big tech" as it existed at the time of the bubble pop trundled along and continued making gobs of money.
Mozilla is probably doomed in the long term. I think they're in the exact same boar as Microsoft, and wholly lack the self-reflection required to turn the ship around.
Firefox will continue to languish while Mozilla execs receive 8-figure bonuses until there's nothing left to extract.
OpenAI and Anthropic are bleeding money and both need hundreds of billions of dollars in the next couple of years to break even.
Oracle is highly overleveraged and I am hoping that the bubble takes them out.
You can find the gory details at Ed Zitron's blog.
https://www.wheresyoured.at/premium-how-the-ai-bubble-bursts...
Anthropic at least has Amazon’s backing. OpenAI is where the industry is stuffing all of its bad debt and transparently bad deals. It’s the sacrificial company this time around.
I would dance on Oracle’s grave, but they have too much staying power because of their core database and ERP business.
Microsoft seem to be pushing all kinds of users away in all directions at the moment while focused on the AI bubble*. Once it bursts/deflats, will they come back?
Or are we looking at a post-Windows future, where MS just focuses on cloud stuff?
(Or will there be a 'we learned from our mistakes, honest' Windows 12 that wins people back in the same way that Win10 did after Win8?)
The problem is, when multi trillion dollars go poof everyone is going to feel the pain, even those that wisely stayed away from investing in the bubble.
And there'll be bullshit bailouts. But I'd rather embrace that pain than living through more cycles of that LLM BS. It will create so many generational problems.
Is it so hard to just read a freaking man page or read a book instead of wasting gallons of water and electricity for some answer that hallucinates command line flags that don't even exist?
You should take the L on this one. Even if you got a "Show HN" post about JAI, I don't think the tech bros are going to be keen on a NSFW AI chatbot website. You still don't have a subscription offering after 2+ years. The users still wonder who in the hell pays to keep the lights on.
The AI bubble will pop in the next year. We are currently in 1998 of the dotcom bubble with another AI winter approaching. LLM and generative AI are this year’s “on the Internet” or “Uber for X” business plans.
Given the direction of the economy, we'd hit a recession and no one will be investing. So a loss for all tech. I argue we already are in one, but the US election cycle wants to delay that reality until 2025.
Though the US did do a pretty big rate cut. I imagine that will at least stall such a bubble burst into 2026 instead.
The economy bubble will be popped post-election (you will know when the Fed starts raising rates again), but CRE (commercial real estate) is likely the catalyst this time.
Within CRE, datacenters are the only item in the green in investors eyes, even more now because of the AI boom. I fully expect that class of investor to then dump more money than ever into energy generation and other sectors related to AI in order to escape the planned crash.
The biggest variable is if the supranational oligarchs are wanting to use this crash to cause a much more major shift in monetary policy such as CBDCs.
The market PE is very high and has been high for a while. Tech and AI are fueling a lot of that. Look what happened to Tesla when the fundamentals started getting a little discouraging. However, I'm not comfortable predicting the "AI bubble" popping next year.
I use it all the time for my QEMU VMs. It’s less hassle than OVS and allows you to configure some complex networks with just a few socket files and tunnels.
I use one VDE switch per network with each network having a single tap interface to my OS bridges. It has been a feature of QEMU since 0.15 and the performance is just as good as taps or OVS.
Charities can hide their percentages when they self-deal with people on their board. It's highly unethical.
Wreaths Across America is notorious for this. All the money they raise goes to the wreath making company that the board members own. When they put out RFPs, as per legal requirements, Worcester Wreath Company is always sole-source provider.
https://popular.info/p/the-truth-about-wreaths-across-americ...
Worcester Wreath Company lost its contract with LL Bean due to their bad busines practices. Wreaths Across America was created by Morril Worcester as a way to sell more wreaths. It's a grift all the way down.
ERAM is multichannel, which is why we do the failover between A-channel and B-channel during APL and OS CUTO. If I remember the SSMs right, we do the update on the B-channel first and once it has been approved by TechOps, A-channel is then updated.
Everything is built to provide a fallback in cause of failure, including the OS updates when they come in.
How are the interstitial ads handled in an OS such as this? Is the ad running time factored into the control system on the kernel level? Do the operators have a realtime safe button press to "skip ad" in high traffic situations?
When privatization is in the news, we like to joke, "I have a Venmo username to send $20 for a practice approach, advise when ready to copy."
Or, "Airlines123, ten miles from POINT, fly heading 360 until established on the localizer, cleared ILS Runway 1 approach. This approach brought to you by <advertiser>."
I've been purchasing older O'Reilly books off of Ebay for the past couple of years. My most recent purcahse was Applying RCS and SCCS by Don Bolinger, which is split between RCS, SCCS, and describing software configuration management in general.
The SCCS interleave format lived in BitKeeper and PVCS/Dimensions.
reply