Big Tech Outspends Venture Capital Firms in AI Investment Frenzy (ft.com) 24
Big tech companies have vastly outspent venture capital groups with investments in generative AI start ups this year, as established giants use their financial muscle to dominate the much-hyped sector. From a report: Microsoft, Google and Amazon last year struck a series of blockbuster deals, amounting to two-thirds of the $27bn raised by fledgling AI companies in 2023, according to new data from private market researchers PitchBook. The huge outlay, which exploded after the launch of OpenAI's ChatGPT in November 2022, highlights how the biggest Silicon Valley groups are crowding out traditional tech investors for the biggest deals in the industry.
The rise of generative AI -- systems capable of producing humanlike video, text, image and audio in seconds -- have also attracted top Silicon Valley investors. But VCs have been outmatched, having been forced to slow down their spending as they adjust to higher interest rates and falling valuations for their portfolio companies.
The rise of generative AI -- systems capable of producing humanlike video, text, image and audio in seconds -- have also attracted top Silicon Valley investors. But VCs have been outmatched, having been forced to slow down their spending as they adjust to higher interest rates and falling valuations for their portfolio companies.
Well, well, well... (Score:2)
that means those who understand risk are clear of those who have cash to spend from their near-monopolies. Peak AI is here, boys, just like peak dot-com was in '99.
Re: (Score:2)
If an AI can't deliver a tub of Ben & Jerry's Cherry Garcia like Kozmo.com could, I'm not interested.
Re: (Score:2)
But what if they can deliver an CRYPTO layer to a currency like beenz or flooz? And what if it has AI BLOCKCHAIN?
Re: (Score:1)
Ben & Jerry's tastes like crap! Buy HÃagen-Dazs ice-cream is you want to taste what real ice cream tastes like [...]
I reckon the days of vi vs. emacs are long over.
Re: (Score:2)
Come on, in the realm of off-the shelf ice cream Norgen Waaz rules.
Re:Well, well, well... (Score:5, Interesting)
In actual reality on the other hand, the problem is supply of money. VCs are funded from wealthy middle and upper class funds when they want to take additional risk for a chance of much higher return in half a decade than they would otherwise get.
Boomer generation is almost finished retiring, which means their risk profile has changed from "maximum profit" to "maximum stability". So VC funding base has crashed, and so has the amount of money they have to lend out.
As opposed to big tech companies that are funded via massive corporate credit lines backed by the company assets.
Re: (Score:2)
You mean "In actual AI reality".
Re: (Score:1)
backed by the company assets
aka the general population. Central banks print money to fund too-big-to fail corporations without any limits
Re: (Score:2)
That's not how monetary system works. At all. It's the commercial banks that fund megacorps, not central banks.
Re: (Score:2)
It helps to follow your own link. And you've trolled me for years, so you know that I do.
https://www.statista.com/stati... [statista.com]
Yup, completely in line with my premise that boomer retiring has recently peaked, when people on the edge of retirement dump everything they can into high risk high return investments and now that we're past the peak, risky capital is on the way down and T-bills and index funds are eating far more of new investments.
Acting in a frenzy is never good (Score:2)
It basically can only work by accident. That is not going to happen here. Anybody sane _knows_ that the AI of the current AI hype is not anywhere as good or capable as claimed and can only deliver some small incremental improvements in a few areas. Nothing at all that would justify the level of investment. Completely irrational "fear of missing out" seems to be the main driver.
Re: (Score:1)
Anybody sane _knows_ that the AI of the current AI hype is not anywhere as good or capable as claimed
I think someone needs to be well-informed and a critical thinker, in addition to sane. The world is full of people eager to Make Money Fa$t on the basis of hype and highly diluted information.
Re: (Score:2)
Yes, probably. I am still getting used to most people not being able to fact-check for shit. Or as Charles Stross put it in a story "The average person understands nothing".
Re: (Score:2)
However, I suspect that less grandiose models can be used to perform more mundane & unimpressive tasks that'd fail to get journalists to write about it in any positive & publicly appealing ways, & that those are the applications that are more efficient, sustainabl
Re: (Score:2)
Agreed. Very specialized LLMs may actually perform, especially when paired with actual deduction engines. For example, an LLM targeted at mathematics that runs everything past Wolfram Alpha (and so merely acts as an interface) may be able to fix the hallucination problem. Of course, it becomes a mere natural language interface at this point and all actual AI is in Wolfram Alpha. Still a valid application and would make the whole more powerful.
But that is a relatively small incremental step, not the "revolut
Re: (Score:2)
Ah yes, the good ol' times come back (Score:4, Interesting)
Remember the late 90s? The dot.com time? When you'd yell "TCP/IP" into a bank and the reply was "how many millions do you need?"
Everyone knew that we're just fleecing the dupes and that the shit ain't going anywhere anytime soon, but hey, if you throw money at me, who am I to say no?
Now the thing to yell into a bank is "AI". Same spiel.
Wonder if there's gonna be Superbowl ads this time around that will crash it. Or whether some spoilsport comes up with something new.
Wonder when the patent wars start... (Score:4, Interesting)
I wonder when the patent wars start, where AI companies stake out claims on various things, perhaps the equivalent of "do this on a phone" claim, and then go at each other with IP violation lawsuits. Assuming patents follow the same path as cellphone technology, which went forward for a while, turned into a patent battle royale, and now little, if any real significant development gets done.
From what I can tell, it seems the AI craze is sort of losing its luster. It has promise, but we have had AI for decades, although not with the evolving LLMs, so for most tasks it is just changing a backend API from what is used right now to someone's LLM and calling it done. Of course, there will be some innovation, perhaps something like what Ives is working on, but this seems lot like the Linux or blockchain days, and with interest rates high, VC capital is a lot hard to come by. Big companies may embrace it, but those are not the place innovation comes from these days, generally, as their DNA is revenue streams and "optimization", not breaking new ground.
Just don't pull *us* into your money orgy (Score:2)
As long as they don't take down the entire economy if and when AI Bubble 2.0 crashes, spend away. (AI bubble 1.0 was in the 80's.)
> where AI companies stake out claims on various things, perhaps the equivalent of "do this on a phone" claim
Seems when a new implementation technique comes out, everyone rushes to convert the prior generation into the next and patent it, as the patent examiners find it "novel" simply because it uses the latest implementation technique, not because it's revolutionary.
Do X by h
Re: (Score:2)
As long as they don't take down the entire economy if and when AI Bubble 2.0 crashes, spend away. (AI bubble 1.0 was in the 80's.)
Don't worry. Investors and the business owners that make the decisions that ultimately lead to full economic spasms will be well protected by making the taxpayers bear the brunt of the inevitable collapse. We must protect the owner class from the calamities of their own making, or else they'd learn a lesson and stop doing stupid shit to further than ambition while risking all of us in the process. And we can't have that. Can we?
Re:Wonder when the patent wars start... (Score:4, Interesting)
The AI patent race is already well underway, with around 20K AI patents already filed.
https://arapackelaw.com/patent... [arapackelaw.com].
I have no doubt that 95% of those patents are...hogwash.