It's Not OpenAI's Sputnik Moment, It's Their Theranos / Elizabeth Holmes Moment
How Much Hype Beyond The AGI Mega Consciousness Is Driving The Billions Funneled To OpenAI and the TechBros?
Remember that weirdo woman in the black turtleneck who spoke in clearly affected yet artificially deep tones while staring at you with unblinking eyes? Her hair was always all messed up because she was working all night at the lab and didn’t have time to shower before the interview. Elizabeth Holmes, touted as the next Steve Jobs because, well, the CNBC vapid investment guru crowd were all impressed that someone would dress like Steve Jobs and talk like a zombie.
Many did, and it turned out that all Theranos demonstrations for investors were faked. The PCR test, while not properly understood and applied, at least works. The Theranos blood tests didn’t work. At all. Holmes was rightfully prosecuted and convicted of fraud.
Holmes was CEO of Theranos, the next biotech revolution in lab testing, it was going to change the world because you could order tests cheaply, and the process was far faster than conventional methods. It was a lie, the processes didn’t work, but it fooled many investors, including former Secretary of State George Schultz and others. The Clintons loved Elizabeth, I don’t know if they fell for her ruse.
I have a theory, and it’s that we have been shown so many stereotypes of personal dysfunction while being told that it’s a sign of genius that we have been conditioned to just accept that the creeps asking for billions just are weirdo grifters. No, because they’re weird, they’re geniuses. They don’t have olfactory senses and are unaware of personal hygiene, but they are exceptionally intelligent. They don’t bathe? Awesome! Here, take my retirement funds.
Sam Altman has an affectation too, I call it the Beta-male Danny-Torrance-Redrum-Mommy-Redrum-My-Little-Friend-Tony voice. But we excuse that because he’s the gatekeeper of even weirder bunch of Tech Bro geniuses who are working on spawning artificial super consciousness. That’s my term, they call it Artificial General Intelligence (AGI), self learning software that can think, and it will give rise to the next round of VC funding for what’s called Artificial Super Intelligence. We need ASI to beat the Chinese, and usher the post scarcity era. I know, this sounds like a cult, I’m not exaggerating.
I’m feeling very irreverent and am drawing clumsy analogies, but bear with me. I think there is a thread here that we are missing amidst the hype of Stargate and the fear of the CCP company DeepSeek stealing secrets and our data. I wrote software for over 30 years, and have experience with all sorts of socially dysfunctional weirdos, as well as some great technicians who knew how to comb their hair. People’s operating systems are emotions. And that’s what we’re dealing with.
The Sputnik Moment, Really?
Recently Trump hosted his first press conference announcing a huge investment in artificial intelligence infrastructure being made by private financiers. Standing with him was Masayoshi Son, CEO of Softbank. Present also was the Walmart version of Bill Gates, Oracle CEO Larry Ellison, and Sam Altman, CEO of OpenA which created the AI app ChatGPT. ChatGPT is the big kahuna burger of AI, with 200 million users on its platform every week.
SoftBank has agreed to supply a $500 billion dollar package to build data centers in the United States with OpenAI, and the rumor is that nuclear reactors could be the power source that will meet the staggering requirements for electricity. Using ChatGPT to perform a search of the Internet takes 10 times more electricity than a plain old Google Search. Take note that the AI summary at the top of the Google search also uses a ton of electricity. So the demand for electricity is going to skyrocket. Special chips called GPUs are needed for the intense processing that it takes for AI to give you that eloquent answer it supplies. Those chips are made primarily by a company called NVIDIA.
On the heels of Trump, Larry and Sam’s side show circus announcement came the news that a Chinese hedge fund had tested a new AI system called DeepSeek that met or exceeded OpenAI’s premier product performance. The shocking fact was they achieved this feat with a different method that cost them only 6 million dollars, and was accomplished with far fewer and less powerful NVIDIA chips. OpenAI does not release many details regarding the cost of “training” their AI systems. However, it is rumored that OpenAI expends $500 million each 6 months to achieve their results when training their AI models how to interpret and answer your queries. You can read more in my previous article.
On Monday, Jan 27 the market was rocked by the single largest drop in NVIDIA stock in it’s history. It lost $600 billion in value on the news that DeepSeek achieved the same results for far less. Additionally, DeepSeek released the software as open source. This means that any company can take the source code and run it on their own, without using DeepSeek computing resources. For free.
In other words, enterprising technologists don’t need billions to compete with OpenAI. They can study the DeepSeek software and perhaps innovate yet again.
On the Monday of NVIDIA’s stock price plummet, Trump stated that this was a wake up call, and that US would have to innovate like we have before. We certainly have the talent to do so.
But OpenAI and the other Tech Bro oligarchs have had the curtain torn away, similar to Elizabeth Holmes. The wake up call is for us who fall for the oligarchs who who offer empty promises and require government action to keep money flowing into there bank accounts.
This has been a Theranos Moment.
If you have found this article enjoyable and compelling, please give it a like. That is gold. It doesn’t cost a thing, but it will help my efforts here. Other platforms will prevent you from simply seeing these articles, and any time they can be surfaced to someone who will take the plunge would be a huge help. Thank you, you’re awesome.
The Future Future State of Round Two of Venture Capitalist Nirvana
What has been revealed is that the amount money that OpenAI claims they need has been called into question. The tests of DeepSeek capability have been confirmed independently, and DeepSeek has published a white paper which describes the process used for building their version of AI. The irony is that OpenAI, while intimating that there has been industrial espionage of their technology, based their process on a white paper published by Google. But that’s not all. OpenAI has been sued for copyright infringement, using protected intellectual property to “train” their AI systems.
There is icing to this cake of irony too. OpenAI used Scarlett Johansen’s lovely voice as the basis of its voice recognition system. Scarlett was outraged and asked them to change it, and miraculously OpenAI did. Currently people are stating that DeepSeek used OpenAI’s AI model via screen scraping and distillation, meaning that they used what they could glean from ChatGPT and trained their AI based on that information. Information that is available through public use of ChatGPT software. It does not denigrate DeepSeek’s innovation. Again, OpenIAI trained their model based on techniques described by Google and by scraping or gleaning from the Internet. The Chinese just trained their system with the OpenAI model.
It makes you wonder why OpenAI hasn’t adopted the same technique, because no one is saying that the CCP stole the OpenAI software. DeepSeek functions because it has been trained by using OpenAI’s response as the foundation, just like OpenAI has used sources of the internet as its foundation. This is NOT stealing intellectual capital from OpenAI.
Aravind Srinivas CEO of Perplexity.AI, a US based company that offers search services combined with AI, explains.
There’s a lot of misconception that China ‘just cloned’ the outputs of OpenAI. This is far from true and reflects incomplete understanding of how these models are trained in the first place…DeepSeek R1 has figured out RL (reinforcement learning) finetuning. They wrote a whole paper on this topic called DeepSeek R1 Zero, where no SFT (supervised fine tuning) was used. And then combined it with some SFT to add domain knowledge with good rejection sampling (aka filtering). The main reason it’s so good is it learned reasoning from scratch rather than imitating other humans or models.
Am I defending the CCP? Nope. Does the CCP steal your shit? You betcha. Don’t go sign up and use the DeepSeek service directly, their terms of service state they will track your keystrokes. You can read about that in detail here.
Did DeekSeek “steal” OpenAI’s work or software. No, they did not.
But since the software that runs that system is available for free, others will soon study and innovate based on that body of work, and run that software cheaply and safely in the US. Those services will compete with OpenAI.
That is a good thing. The Tech Bros are oligarchs, and we need to make sure that they don’t shield themselves in a Too Big To Fail 2.0 moment while claiming that it’s a matter of national security that we protect their bloated expensive processes. Under the Biden administration, the National Security Memorandum declared that AI was a primary factor in our nation’s defense. In conjunction with Executive Order 14110, AI was to be governed to ensure safety, and very onerous provisions would be applied to the Ai sector of the tech industry. This meant that smaller competitors would not survive given the additional steps for ensuring that their AI software met government standards that OpenAI could easily fulfill. Sam Altman welcomed the NSM.
Of course he did, it created a near monopoly for OpenAI. Soon after the NSM was announced, OpenAI was awarded a DoD contract in concert with Palentir. Upon taking office, Trump cancelled the NSM.
The Therano Moment is this: OpenAI has yet to turn a profit, and while the software does generate and analyze text in an impressive fashion, there is a huge amount of hype around AI becoming conscious and replacing white collar workers. According to OpenAi and industry pundits, more processing power, and therefore money, is needed to achieve this stunning milestone. When you go on YouTube you see countless videos of people chanting about the emergence of sentient artificial consciousness. Data from Star Trek, or T2 Terminators. It’s just around the corner. Altman makes statements that AGI it’s near, perhaps within 1000 days. It’s so confusing because sometimes the tech pundits uses AGI, others us ASI. But the whatever it is to be better than human is almost here.
That hype makes people get out their checkbooks. That hype gets jealous Tech Bro wannabes like Larry Ellison to join the cool kids on stage and spew CEO bromide word salad. We’ll cure cancer, OpenAI will let you lease a PhD level of super intelligence so you can avoid H1B hiring like Musk was screaming about a few weeks ago. Sorry Elawn. And perhaps with AGI / ASI super intelligence you can fire all the annoying laptop class members who just share memes all day on social media. Hurry, the AGI sale ends soon.
And now the CCP may achieve it before the US does. Remember, people’s operating system is emotion.
I have to thank Demi of
who indulges me trading memes. She got me thinking about Danny Torrance and his friend “Tony”, that gravelly voice he used to channel spirits.
The universal bromide from that video "robots don't get sick or need lunch breaks" is laughable. They get "sick" when a servo fails or mechanical pieces snap from metal fatigue, or even when their code gets corrupted by a lightning EMP etc. Also, not only do they need "lunch", they constantly need fed electricity, if their diet is interrupted for a nanosecond they either quit or go berserk. There is much more to reflect upon with your excellent post, but I haven't sufficient coffee in my sentient brain to sort out.
Anyway we have a beautiful soft lovely snow outside, time for mind to take a break 😬