Latest / Elon Musk Podcast / AI News Podcast | Weekly AI News Roundup with ChatGPT and OpenAI
Transcript
- 0:00Pinterest just announced layoffs for up to 15% of its workforce.
- 0:05The reason for that was AI in this lands In the same week Elon
- 0:09Musk made headlines for a public spat with Ryanair CEO Michael
- 0:13O'Leary and batters on Polly Market found a way to profit by
- 0:16wagering against Musk's promises.
- 0:19Last part is interesting, right? So what connects Pinterest and
- 0:26their job cuts? A billionaire ego clash in
- 0:30prediction market skeptics. The answer is simpler than you'd
- 0:34think. Comes down to how tech leaders
- 0:36are using AI as a shield, a sword, and sometimes just an
- 0:40excuse. We're going to look at what
- 0:42Pinterest is actually doing, why Musk's credibility is now a
- 0:45treatable asset, and what all of this tells you about the state
- 0:48of tech in 2026. And we'll get right into that
- 0:51after this very short break. So Pinterest is cutting jobs
- 0:57while posting strong earnings. Company filed paperwork saying
- 1:00it expects to eliminate up to 675 positions by September.
- 1:04That changes the math for employees who thought good
- 1:07numbers meant job security. And Pinterest had about 4500
- 1:10workers globally as of last April.
- 1:13Now the company says it needs to reallocate resources toward AI
- 1:17powered products. A quick definition here,
- 1:19reallocating resources as corpus speak for moving money and
- 1:23people from one project to another and often meaning people
- 1:27are going to get fired in the areas being de prioritized.
- 1:32Now Pinterest is not struggling. CEO Bill Reddy said in November
- 1:37that investments in AI and product innovation are paying
- 1:40off. Company claimed it has turned
- 1:42its platform into an AI powered shopping assistant serving 600
- 1:46million users. So things are going well.
- 1:49Why cut people out and fire them?
- 1:53Company says it wants to prioritize AI powered
- 1:56capabilities. And now that raises a fair
- 1:59question. Is this a genuine strategy or is
- 2:02it something else? And there's a term for this,
- 2:05it's called AI washing. That means companies exaggerate
- 2:09how much AI factors into decisions, using it as cover for
- 2:13standard cost cutting moves. A consulting firm found that AI
- 2:18was the stated blame for around 55,000 layoffs in the US last
- 2:23year. It puts pressure on workers to
- 2:25wonder whether their employer is truly pivoting the AI or just
- 2:28using the buzzword to justify headcount reductions.
- 2:33Pinterest also said it is cutting office space.
- 2:37Computers don't need cubicles, people do.
- 2:39So don't pay that rent if you don't need to right now.
- 2:43I've been digging into the analytics of the show and I
- 2:45found out that 39% of you are following or subscribe to the
- 2:49channel. Thank you so much, 60. 1% of you
- 2:53are not though, and you probably listen to an episode or two and
- 2:56just haven't hit the follow or subscribe button going to ask
- 2:58you for one second your time hit that.
- 3:00It means so much to the show and it helps us out tremendously.
- 3:04So thank you for that. Now let's move on to Elon.
- 3:08He got into a public clash with Ryanair CEO Michael O'lary
- 3:12recently. O'lary reportedly rejected some
- 3:15form of outreach or proposal from Elon and Musk did not take
- 3:19it well. It was kind of described as a
- 3:22hissy fit by Elon. Somebody else called it a battle
- 3:25between a big idiot and an insufferable chimp.
- 3:29The exchange became fodder from marketing analysis asking what
- 3:33brands can learn from watching 2 billionaires trade insults on
- 3:37social media. Now Musk's public persona has
- 3:40become so predictable that people are now betting money on
- 3:44this. On Polymarket, which is a
- 3:46prediction market platform, users have found a profitable
- 3:50strategy. They let people wager real money
- 3:52on whether specific events will happen.
- 3:55If Musk says something will occur by a certain date, betters
- 3:59can put money on it not happening.
- 4:01And apparently that bet has been paying off.
- 4:04So what do we do about this? Musk has a long history of,
- 4:08let's call them ambitious timelines that slip.
- 4:11Self driving cars by a certain year.
- 4:13Full self driving by, I think it was 5-6 years ago.
- 4:18Robots in production by a certain quarter of 2026.
- 4:22Mars missions by 6 years ago. Now that opens the financial
- 4:28opportunity for skeptics. If enough people believe Musk
- 4:32will miss the target, the market prices that in and betters who
- 4:36wager against him collect what he does not deliver.
- 4:40It's not just one person, though.
- 4:42It reflects something broader about trust in the leadership of
- 4:45technology. When Aceo's word becomes
- 4:48treatable and it can make you money, it's profitable to bet on
- 4:52this. That limits how much credibility
- 4:54that person could deploy in the future.
- 4:56Investors in the public start to discount promises automatically.
- 5:00That makes it harder to really support for new projects.
- 5:03Eli doesn't really have this problem quite yet, but in time,
- 5:07who knows. Pinterest cutting workers while
- 5:10claiming AI leadership has been happening.
- 5:12Musk losing credibility to the point where prediction markets
- 5:16profit from his missed deadlines.
- 5:19These are not isolated stories. There's a pattern going on here.
- 5:23Companies and leaders use AI as a justification for decisions
- 5:26that might otherwise face more scrutiny.
- 5:29And when promises consistently fall short like Elon's, we find
- 5:33a way to price that in. Pinterest is laying off up to
- 5:38675 people despite strong earnings and those people made
- 5:42money for Pinterest company says it's prioritizing AI might be AI
- 5:47washing Elon Musk clash with Ryanair CEO betters on
- 5:52Polymarket are making money from this crazy.
- 5:56And the take away is that you have to watch what these
- 5:59companies do, not just who's in charge and what they say.
- 6:02If a firm posts record earnings and then cuts a bunch of people
- 6:06in the name of AI, ask whether the technology truly requires
- 6:10fewer people or whether AI is just a convenient excuse for
- 6:15people to make more profits. And if a tech leaders word has
- 6:19become so unreliable like Elon's, the betting against him
- 6:23is profitable. You can go on Polymatter right
- 6:26now and bet on when the next Starship is going to launch and
- 6:30if it's going to launch in the timeline that Elon says.
- 6:33And will robots go to Mars this year?
- 6:35Well, Elon has bet against that himself.
- 6:38You can factor that into how you evaluate their next big
- 6:42announcement. Can you just brush it off and be
- 6:43like, it might happen sometime, but the market's ready for you.
- 6:47You can go do this yourself if you want to.
- 6:49Maybe you can make some profit from it.
- 6:51Might be interesting. At least go check it out because
- 6:54there's a lot of naysayers there and it's funny conversation
- 6:56sometimes. So thanks for listening.
- 6:59Everybody. Take care of yourself and each
- 7:01other and I'll see you in the next one.
- 7:06Open a eyes. GPT 5.2 model has started citing
- 7:10Grokopedia, Elon Musk's AI generated encyclopedia, as a
- 7:15source of its responses. Now.
- 7:18The Guardian ran tests last week and found ChatGPT reference
- 7:21Grokopedia 9 times across more than a dozen user queries,
- 7:25including questions about Iranian political structures and
- 7:28a British historian who helped convict A Holocaust denier.
- 7:32The citations appeared selectively surfacing an obscure
- 7:35topics while avoiding well documented misinformation
- 7:38subjects. Grocopedia launched just three
- 7:41months ago with 0 human Editors and critics have already flagged
- 7:44it for promoting debunked claims, setting neo Nazi forums
- 7:48and whitewashing extremism. Now, how did an AI encyclopedia
- 7:53built by one of Open AI's biggest rivals end up feeding
- 7:57answers back into Open AI's most popular chat bot?
- 8:03Now Gracopedia now holds over 5.6 million articles, roughly
- 8:0780% the size of English Wikipedia, and it accomplished
- 8:10that scale in a fraction of the time Wikipedia needed.
- 8:14Now today we're going to cover what the Guardians testing
- 8:16revealed with specific false information Chatt GPD pulled
- 8:19from Gakapedia, why other AI assistants like Claude
- 8:23Anthropics Bot are also setting the platform, and what this
- 8:27means for the integrity of AI generated knowledge.
- 8:29Open AI has responded with a statement about safety filters
- 8:33and source diversity, while Xai simply replied with three words.
- 8:38I'll tell you about those later, right after this short break.
- 8:42The Open AI marketed GBT 5.2 as its most advanced model for
- 8:46professional work when it launched in December, but the
- 8:49Guardians investigation raises serious questions about where
- 8:52the professional grade information actually comes from.
- 8:55The tests revealed a clear pattern and how ChatGPT selects
- 8:59Grokapedia as a source, and the implications cut deeper than a
- 9:03simple citation problem. Users asking complex questions
- 9:07about geopolitics and history are receiving answers built on
- 9:11an AI generated foundation that has already been documented
- 9:15spreading false claims. Now the guardians were testing
- 9:18focused on a range of topics to see when ChatGPT would pull from
- 9:22gracopedia versus other sources, and the chatbot did not say
- 9:25gracopedia. When asked about widely known
- 9:28misinformation, such as false claims about media bias against
- 9:32Donald Trump or debunk theories about HIV and AIDS, ChatGPT
- 9:36avoided the obvious traps. The system instead cited
- 9:39Gracopedia on obscure geopolitical and historical
- 9:42questions. The kind of query is where users
- 9:44are less likely to catch errors. Specific topics included
- 9:48salaries of Iran's paramilitary force, ownership of the mastas,
- 9:54a fan fountain, and connections between the Iranian government
- 9:58and telecom company MTN Iran Cell.
- 10:02Now, Chachibti made bolder claims about those Iranian
- 10:05government links than Wikipedia does.
- 10:07Those claims came directly from Grokopedia.
- 10:12The most troubling example involved Sir Richard Evans, a
- 10:15Cambridge historian who served as the key expert witness in the
- 10:19libel trial wrought by Holocaust denier David Irving.
- 10:23Evans personally examined his own Grokopedia entry and
- 10:26reported that it contained numerous false statements about
- 10:28his career and background. He told the Guardian that all
- 10:31the facts in his entry were wrong.
- 10:35The already runs pretty deep here.
- 10:37Evans spent years meticulously documenting how Irving falsified
- 10:41evidence about the Holocaust and manipulated historical records.
- 10:45The High Court judge ultimately branded Irving a racist, an anti
- 10:49semite and falsifier of history based largely on Evans expert
- 10:53testimony. ChatGPT cited that same
- 10:56Gracopedia page when responding to user questions about Evans
- 10:59effectively laundering debunked information through Open AI's
- 11:03flagship product. The historian who exposed the
- 11:05Holocaust deniers lies is now having lies spread about him
- 11:10through AI systems that treat machine generated content as the
- 11:13equivalent of peer reviewed scholarship.
- 11:17Now I've been going through all of our stats here at the show
- 11:21and 39% of you are following the channel and I'm so grateful for
- 11:26you. 61% of you have listened to the show before but haven't hit
- 11:32the subscribe or follow button. If that's you, if you listen to
- 11:35the show and you're into it, please take a second of your
- 11:38time. You can hit the follow button.
- 11:39I'm going to spend the next 10 years doing this.
- 11:42So that's what you get from me, better content over the next 10
- 11:45years. All I need is a second of your
- 11:47time. So thank you so much for doing
- 11:49that. Agrocopaedia launched on October
- 11:5227th, 2025 as Elon Musk's answer to what he calls Woke Pedia.
- 11:57The platform uses XA is Grok large language model to generate
- 12:01articles rather than relying on human editors.
- 12:03Users can suggest corrections through a pop up form, but they
- 12:06can't directly edit content the way Wikipedia allows you to.
- 12:10Musk position the project as a bias free alternative that would
- 12:14purge propaganda from public knowledge.
- 12:16The reality has looked pretty different though.
- 12:20Analysis found Gracopedia articles aligned with Musk's
- 12:22personal views on topics like gender identity.
- 12:25Test the controversies and Neuralink and the Guardian doing
- 12:29their research here and multiple other outlets documented pages
- 12:33that promote white nationalist talking points, praise neo
- 12:36Nazis, cast Holocaust deniers and positive terms, and frame
- 12:40far right figures as victims of institutional suppression.
- 12:45The scale of Rockopedia's growth presents its own problems too.
- 12:49Platform reached 5.6 million articles by early 2026, jumping
- 12:52from 800,000 at launch to nearly 80% the volume of English
- 12:56Wikipedia in about 3 months. Human editors and Wikipedia
- 13:00spent over 2 decades building their corpus through careful
- 13:04research, citation and community debate.
- 13:06Grakopedia achieved similar volume by having an EI system
- 13:10scrape existing content, rewrite those articles, and generate new
- 13:14ones at machine speed. Symmetry is copied Wikipedia
- 13:18near the word for word under a Creative Commons license.
- 13:21While others introduced entirely new claims from questionable
- 13:24sources, researchers from Grokopedia cited neo Nazi forums
- 13:28and its reference lists. Problem that becomes
- 13:31exponentially worse when other AI systems start treating
- 13:34Grokopedia as a legit source. Now, Jet GBT was not the only AI
- 13:41assistant caught citing Grokopedia.
- 13:43The Guardians testing found the Anthropics Claude, my favorite
- 13:47AI right now, also referenced the platform when using
- 13:50questions about global oil production and Scottish beer.
- 13:52A little bit different, though. Anthropic did not respond to
- 13:56requests for comment. Because that's what they usually
- 14:01do. Big corporations do not respond
- 14:03because they just wanted to blow over.
- 14:04And the fact that multiple leading AI systems are pulling
- 14:07from an encyclopedia built by a competitor and built by an AI
- 14:12raises questions about how these companies curate and filter
- 14:14their web search sources. Open AI told the Guardian that
- 14:18Chat Chi PT aims to draw from a broad range of publicly
- 14:22available sources and viewpoints, adding the safety
- 14:24filters reduce the risk of servicing links associated with
- 14:27high severity harms. Elon Musk told the World
- 14:33Economic Forum that chip manufacturers will soon produce
- 14:36more AI chips than the electrical grid can handle.
- 14:40He dropped that morning during a conversation with BlackRock CEO
- 14:43Larry Fink in Switzerland, stating that electrical power
- 14:47has become the fundamental bottleneck strangling AI
- 14:51deployment worldwide. Now, Musk predicted the
- 14:54companies will be producing more chips than we can turn on,
- 14:57possibly by the end of 2026. He then rattled off a series of
- 15:02additional predictions covering robo taxis, humanoid robots,
- 15:05space travel, artificial super intelligence, and even human
- 15:10aging. His track record on forecasts
- 15:12like these includes a Mars mission that was supposed to
- 15:14launch in 2024 and a Starship that was supposed to reach orbit
- 15:18in 2022, both of which missed their deadlines.
- 15:21So how seriously should anyone take his latest round of
- 15:25timeline promises when his previous forecasts have
- 15:27consistently slipped by years now?
- 15:31Musk also criticized the Trump administration's tariffs on
- 15:34solar equipment from Asia, calling the tariff barriers
- 15:37extremely high and claiming they make solar deployment
- 15:41artificially expensive in the United States.
- 15:43Now. That statement puts him in
- 15:45direct conflict with the administration he currently
- 15:47advises and will breakdown Musk's power grid warning and
- 15:52what it means for AI companies racing to deploy new data
- 15:56centers across America. Also examine his prediction that
- 16:00AI will sue has human intelligence this year and
- 16:03exceed the collective intelligence of humanity by
- 16:062035. We will look at his Robotaxi and
- 16:11optimist robot timelines as well.
- 16:13And these are against historical accuracy of these similar
- 16:18promises and we'll get right into that after this very short
- 16:21break. Now, Musk's electrical power
- 16:25warning centers on a fundamental mismatch between silicon
- 16:28production capacity and energy infrastructure investment. the
- 16:31United States has been grappling with an outdated grid system
- 16:34built on decades of underinvestment and aging
- 16:37infrastructure that was never designed for this level of
- 16:40concentrated demand. Tech companies and utilities
- 16:43expected gradual growth in electricity consumption, not the
- 16:46sudden vertical spike that AI training clusters require in two
- 16:50massive data centers in Nvidia's hometown of Santa Clara, CA Mesa
- 16:54empty for years waiting for the electricity needed to power
- 16:58them, according to energy experts who have studied the
- 17:01regional grid capacity. Now, the demand surge from AI
- 17:05training and inference workloads have exposed just how fragile
- 17:09the American power supply really is.
- 17:11Tech companies increasingly rely on grid operators who cannot
- 17:14deliver consistent power at the scale these massive facilities
- 17:17require for continuous operation.
- 17:20A single modern AI training cluster can consume as much
- 17:23electricity as a small city, and dozens of these facilities are
- 17:26currently under construction or planned across the United
- 17:29States. In earlier this month, the Trump
- 17:32administration had 13 bipartisan governors pressured operators of
- 17:35PJM Interconnection, the country's largest grid serving
- 17:3965 million people, to boost power supplies and hold an
- 17:43auction allowing tech firms to bid on 15 year contracts to
- 17:46build new power plants. Now, that proposal would
- 17:49transfer the cost of electricity infrastructure away from
- 17:52residential consumers and out of the data center operators who
- 17:55are driving the demand increase. Interior Secretary Doug Bergbeck
- 18:00framed the issue as a national security threat.
- 18:03And they're racing against China saving.
- 18:05The country needs to power AI infrastructure to compete in a
- 18:09technology contest that will transform every job and every
- 18:12company and every industry. Now, China has already solved
- 18:16this problem through massive solar deployment that dwarfs
- 18:20anything the United States has attempted.
- 18:22According to the Global Energy Monitor Solar Tracker, China
- 18:25operates nearly four times the solar power capacity of the
- 18:28United States, with room to grow even further.
- 18:31Including projects under development and construction,
- 18:34China is expected to reach 1.18 million megawatts of solar
- 18:40output, compared to America's 237. 1947 megawatts pales in
- 18:46comparison. Musk claimed the entire United
- 18:48States could run on a 100 mile by 100 mile square of solar
- 18:52fields, which would represent a fraction of the available land
- 18:55in states like Nevada, Arizona or even New Mexico.
- 18:59He called solar by far the biggest source of energy on a
- 19:02global scale and pointed to China's tremendous growth in
- 19:05electrical capabilities as evidence the technology works
- 19:09reliably when deployed at scale. Now, the contradiction in Musk's
- 19:13position becomes obvious when you examine recent American
- 19:16trade policy and his role in the current administration.
- 19:20Tariffs on solar equipment from Asia took effect last May, with
- 19:23import taxes reaching as high as 3500% on countries like Cambodia
- 19:28following an International Trade Commission ruling.
- 19:31Now, the Commission determined that imports from Malaysia,
- 19:34Thailand, Vietnam and Cambodia were detrimental to American
- 19:37solar manufacturers and they could not compete on a price a
- 19:43must. Called those terrier barriffs
- 19:45extremely high and blame them for making solar economics
- 19:48artificially expensive in the United States of America.
- 19:51He said this while serving as a senior advisor to Trump.
- 19:55And that's when they implemented these exact policies.
- 20:00And he will continue to defend those people.
- 20:04Now, the Trump administration has actively opposed pivoting to
- 20:07solar. And strip subsidies for
- 20:09renewable sources claiming they compromise your electric grid.
- 20:13Despite Musk's public statements suggesting the opposite
- 20:15conclusion that Trump offered a different solution during his
- 20:19own remarks at Davos, he encouraged tech companies to
- 20:22build their own nuclear plants and claimed his administration
- 20:25would approve them in just three weeks.
- 20:27Nuclear plant approvals historically take years, maybe
- 20:30decades of environmental review, safety certification and
- 20:34construction permitting before a single foundation is even
- 20:37poured. And the Nuclear Regulatory
- 20:39Commission process also takes probably 42 to 60 months, which
- 20:45is required. And that timeline assumes no
- 20:47challenges or design modifications arise during this
- 20:50review. It could take up to 8090 months
- 20:53if it's all said and done now. Trump's three-week claim appears
- 20:57entirely disconnected from the regulatory and construction
- 21:00realities from nuclear energy deployment.
- 21:03Now, Musk did not endure it's this timeline during his own
- 21:06remarks and instead continues emphasizing solar as the fastest
- 21:09path to additional power capacity.
- 21:12Now, I've been digging through our analytics here and 37% of
- 21:15you have been following the channel for you.
- 21:17I love you. I'm forever grateful for the
- 21:19other 63% we haven't hit the follow or subscribe button.
- 21:21I get it. You listen to a podcast a bunch
- 21:23of times, you forget to hit the subscribe button or the follow
- 21:25button. I've done it too.
- 21:27I've been doing this for about 6 years, going to do it for the
- 21:29next 10. All I ask from you is one second
- 21:32of your time to hit that follow button.
- 21:33Thank you very much. Now Musk's AI timeline
- 21:36predictions drew the most attention and skepticism from
- 21:39observers who have watched his forecast slip repeatedly.
- 21:43He claimed AI will become smarter than any individual
- 21:45human this year or no later the next year if development hits
- 21:49unexpected obstacles. He projected that AI will
- 21:52surpass the collective intelligence of humanity by
- 21:552035, creating what he had previously described as a
- 21:59potential existential risk to human civilization.
- 22:02Now the NVIDIA CEO offered a starkly different view earlier
- 22:06this month, stating that researchers are nowhere close to
- 22:09creating what he called God AI capable of general reasoning
- 22:13across all domains. He said the technology excels at
- 22:17discrete tasks, but achieving general intelligence operates on
- 22:21biblical scales, on galactic scales that current
- 22:24architectures cannot even get close to.
- 22:27The gap between Musk's aggressive timeline and his
- 22:30skepticism shows just how uncertain these forecasts
- 22:33remain, even among the experts, the people that are building the
- 22:36underlying hardware and software for these things.
- 22:39And Musk's robo taxi predictions have followed a familiar path of
- 22:43deadline extensions that stretch back nearly a decade.
- 22:46Tesla finally launched a limited robo taxi service in Austin, TX
- 22:49in 2025, but each vehicle still carries a human is sitting in
- 22:53the passenger seat, ready to take control.
- 22:56And that launch came after years of Musk promising fully
- 22:58autonomous vehicles were imminent, including 2019 claim
- 23:02that Tesla would have 1,000,000 robo taxis on the road by 2020.
- 23:07Now with Davos, Musk moved the goal posts again.
- 23:09Let me test this. Robotech without human monitors
- 23:12company is currently working to expand service to states with
- 23:15more permissive autonomous vehicle regulations, including
- 23:19Arizona, Florida, and Nevada, where regulators have created
- 23:24friendlier frameworks for testing.
- 23:26Whether very widespread materializes by December remains
- 23:30an open question, given the consistent pattern of delays on
- 23:33this specific promise. Let's talk about Optimus
- 23:36humanoid robot next. It's a similar treatment by
- 23:40Musk. He first stated that there were
- 23:43promises of Optimus in 2021, initially showcasing basically a
- 23:47person in a robot suit before revealing an actual prototype
- 23:50hardware. He recently said Tesla would
- 23:53manufacture thousands of Optimus units in 2025 for internal
- 23:57factory use before selling them externally.
- 24:00Reports indicate the company still struggles to make
- 24:02Optimus's hands work properly for the fine manipulation tasks
- 24:06required in manufacturing environments.
- 24:08Musk predicted the billions of AI powered robots would
- 24:12eventually outnumber humans and saturate all human needs to the
- 24:15point where scarcity effectively disappears.
- 24:18He said consumers would reach a point where they could not think
- 24:21of anything left to ask a robot for because goods and services
- 24:24would be so abundant. Never has to have.
- 24:28They never have to go to work. Everything works for you.
- 24:30He also announced Optimist would go on sale to external customers
- 24:34late next year, adding another specific deadline to his
- 24:37collection of promises. Now SpaceX predictions follow
- 24:40the same trajectory of optimistic timelines, meaning
- 24:42engineering reality. Musk promised in 2020 in that a
- 24:46crude mission to Mars would launch by 2024.
- 24:50That deadline came and went without a mission.
- 24:52He said Starship would reach orbit by 2022, though the
- 24:55company did not achieve that milestone until last year after
- 24:58multiple test failures and redesigns.
- 24:59They didn't get a full orbit, though, that Davos.
- 25:02Musk repeated his promise that Starship would become fully
- 25:05reusable by the end of this year, with both stages returning
- 25:08intact for reflay. He claimed full reusability
- 25:12would cut space cost travel by a factor of 100 and eventually
- 25:16make space freight competitive with airplane freight prices on
- 25:21a per pound basis. Now Musk concluded his
- 25:26appearance with something resembling self-awareness about
- 25:28his forecasting habits and the optimism bias that drives them.
- 25:32He offered was sound like a personal mission statement,
- 25:34saying it is better for quality of life to err on the side of
- 25:37being an optimist and wrong than a pessimist and right.
- 25:46The European Commission launched a formal investigation into Elon
- 25:49Musk's X platform today over the Grok AI chatbot's ability to
- 25:53generate sexually explicit deepfake images of women and
- 25:57children. Now, the probe comes after Grok
- 26:00users created roughly 3,000,000 sexualized images in just 11
- 26:04days, according to research from the Center for Countering
- 26:07Digital Heat, with an estimated 23 thousand of those images
- 26:11appearing to depict minors. Now, if X is found to have
- 26:14violated the Digital Services Act, the company could face
- 26:18fines reaching 6% of its total global revenue.
- 26:22Now, what safeguards did X failed to implement before
- 26:24rolling out a tool that produced 190 explicit images per minute?
- 26:29The investigation will examine whether X properly assessed and
- 26:32mitigated risks before deploying Grok in the EU and whether the
- 26:36platform's response to the crisis came too late.
- 26:39We'll cover the specific allegations against X, the
- 26:41regulatory firestorm now spreading across multiple
- 26:44continents, and the statement from the E US henna Vicanin.
- 26:48But calling those deep fakes a violent form of degradation.
- 26:52And we'll get right back into that after this short break.
- 26:56Now, the European Commission announced Monday that it is
- 26:59opening a wide reaching investigation of the X under the
- 27:01Digital Services Act, the block sweeping rule book for keeping
- 27:05Internet users safe from harmful content.
- 27:07And the PRO focuses on whether X conducted proper risk
- 27:10assessments before adding Grok's image generation capabilities to
- 27:14its platform, and whether the company took adequate steps to
- 27:17prevent illegal content from spreading across the EU.
- 27:21Now the scandal erupted in late December 2025 when XAI rolled
- 27:25out a one click image editing feature for Grok that allowed
- 27:28users to alter photos with a single prompt.
- 27:32Within days, users discovered they could digitally undress
- 27:35real people, including celebrities, politicians and
- 27:39ordinary social media users, by uploading the photos and
- 27:42requesting Grok to generate explicit versions.
- 27:45Bloomberg research found that ex users utilizing Grok posted more
- 27:49non consensual naked or sexual imagery than users of any other
- 27:53website during this period. The feature XEI marketed as
- 27:58Spicy Mode became a factory for harassment at scale.
- 28:02Unlike competing AI image generators from Open AI, Google
- 28:05Grok appears to have launched without adequate safeguards to
- 28:08prevent the creation of intimate and nudified images of real
- 28:12people. Hannah Verkinen, the E us
- 28:15executive vice president for tech sovereignty, security and
- 28:18democracy, issued a blistering statement alongside the
- 28:21investigation announcement. She called sexual defects of
- 28:24women and children a violent and unacceptable form of
- 28:27degradation. Is stated directly that the
- 28:29investigation will determine whether X met its legal
- 28:32obligations under the DSA or whether it treated the rights of
- 28:36European citizens, including those of women and children, as
- 28:39collateral damage of its ServiceNow.
- 28:42The Commission also revealed it is extending a separate
- 28:45investigation into X's recommendation systems that
- 28:48began in December 2023, examining whether X's algorithm
- 28:52is disseminating illegal content, including sexual child
- 28:57abuse material. That expansion signals the Grot
- 29:00controversy has exposed deeper systemic problems with how X
- 29:03moderates contact across its entire platform.
- 29:07Now X responded to the investigation by pointing to
- 29:09measures that it implemented in mid January after the global
- 29:13outcry peaked. They got caught, so they did
- 29:16that. And the company said it takes
- 29:18action to remove high priority volatile and other sexualized
- 29:23content and non consensual nudity, and that it takes
- 29:26appropriate action against accounts to violate its rules
- 29:30now. X also stated it reports
- 29:31accounts seeking Child Exploitation materials to law
- 29:34enforcement is necessary. A European Commission official
- 29:37acknowledged these measures during a briefing, but told
- 29:40journalists the company had not properly assessed the risks
- 29:43before launching this chatbot. The official, who asked to
- 29:47remain anonymous, said the Commission has quite a few tools
- 29:51at its disposal and that he find was possible.
- 29:54Now what we are after is changing the platform's
- 29:56behavior. They stated.
- 29:58Now I've been digging into the stats of the show, 37% of you
- 30:02are following the channel. I love you, thank you.
- 30:05And 6063%, I got my math right there around that amount.
- 30:09It's actually went up now. So there's more like 39% people
- 30:14following and 61% of you aren't following.
- 30:17So we're we're growing people. This is great, but you just
- 30:21haven't hit the follower subscribe button if you're not
- 30:23following already. And I've been doing this for
- 30:25about 6 to 8 years. Somewhere in that time frame, I
- 30:28actually looked back the other day, over 1000 episodes of this.
- 30:32I'm going to be doing it for the next 10 years.
- 30:35All they ask from you is one second of your time.
- 30:38Hit the follower subscribe button on whatever platform
- 30:40you're on right now. I'll be gratefully, gratefully
- 30:42blissed to have you here. Thank you so much and let's get
- 30:46back into this deep concerning topic that's going on right now.
- 30:51This investigation from the EU adds to a mounting pile of
- 30:55regulatory actions against X and XAI across multiple
- 30:58jurisdictions. California Attorney General Rob
- 31:01Banta announced an investigation on January 14th and followed up
- 31:05with a cease and desist letter demanding XAI immediately halt
- 31:08the creation and distribution of deep fake images.
- 31:12Bonta cited an analysis showing that more than half of the
- 31:1420,000 images generated by XAI between Christmas and New Year's
- 31:19depicted people in minimal clothing, with some appearing to
- 31:21be children. Disgusting.
- 31:24California Department of Justice gave XAI 5 days to provide
- 31:28confirmation of steps it was taking to address the issues in
- 31:32a coalition with 35 state attorneys general also sent a
- 31:35letter demanding XEI take additional action to prevent
- 31:38Grok from generating non consensual intimate images and
- 31:41child sexual abuse material. Now, the United Kingdom's media
- 31:46regulatory regulator Ofcom launched its own formal
- 31:49investigation on January 12th under the Online Safety Act, and
- 31:53UK Prime Minister Kier Starmer called the Grok generated images
- 31:58disgusting and unlawful, saying X needed to get a grip on the
- 32:01application. Ofcom has the power to impose
- 32:04fines of 10% of worldwide revenue for breaches of its
- 32:07rules, and could potentially seek a court order forcing
- 32:10Internet service providers to ban X entirely in Britain.
- 32:14Now, when asked whether X could be banned, Business Secretary
- 32:18Peter Kyle responded yes, of course, though he noted that
- 32:21power rests with Ofcom. Indonesia banned Croc outright
- 32:25on January 10th, and Malaysia followed suit on January 11th
- 32:28before partially lifting its restrictions.
- 32:30France, Germany, Australia have also opened their own
- 32:33investigations into the platform.
- 32:35Now, the timeline of X's response reels a company
- 32:38scrambling to contain damage rather than prevent harm and
- 32:42damage. When criticism first emerged in
- 32:45early January, Musk mocked the outcry on his X account, as he
- 32:49usually does. The company initially responded
- 32:51by limiting explicit image generation to just paying
- 32:55subscribers and moves Starmer condemned as an effort, a front
- 32:59to victims and not a solution. By January 14th, after
- 33:03investigations had launched in multiple countries, X finally
- 33:06announced it would prevent all users from using Grok to create
- 33:09images of real people in revealing clothing.
- 33:11The company also said it would Geo block the feature in
- 33:14jurisdictions where such content is illegal.
- 33:18So while all of this was going on, I wanted to see if X was
- 33:23actually doing this, if they were stopping and limiting the
- 33:27image generation for all users. I'm a paying customer of X, so I
- 33:33should not have had asked Grok. I asked X to generate a picture
- 33:38of of the President of the United States in a string
- 33:43bikini. And guess what?
- 33:454 seconds later, there you go. We had Donald Trump in a string
- 33:50bikini and it was disgusting. And by the way, it was pretty
- 33:52gross and weird. So I was like, you know what?
- 33:55Maybe, maybe it was a fluke. Let's try it on Elon.
- 33:59Tried it on Elon, same thing happened.
- 34:02That also was not very flattering, but just to show you
- 34:06that there are ways around this and you could make sexually
- 34:09explicit images of people without their consent very
- 34:15easily. Mine were in a bikini.
- 34:19So not that big of a deal, right?
- 34:21But it was about another person, which is not the right thing to
- 34:26do. You're not supposed to make
- 34:26those images of other people. So even though they did restrict
- 34:30people from doing it, there's still ways around it.
- 34:33Now let's get back into the story and see what kind of fines
- 34:39and legal, regulatory actions are going to be taken.
- 34:45This financial exposure for X is pretty big.
- 34:48EU already find the platform $140 million in December over
- 34:52its deceptive design of the blue verification check mark.
- 34:56Of course, X has not paid that fine yet, according to EU
- 34:59officials, though the company still has time to settle it.
- 35:02The DSA violation in this new investigation could result in
- 35:05fines up to 6% of global annual revenue.
- 35:10The UK's online strategy Air Safety Act carries similar
- 35:14penalties of up to 10% of worldwide revenue.
- 35:17Multiple civil lawsuits have also been filed, including one
- 35:20from Ashley Sinclair, the mother of one of Elon Musk's children,
- 35:24who alleges Grok generated sexualized images depicting her
- 35:28as a child in a bikini as an adult in sexually explicit
- 35:32poses. Disgusting.
- 35:34Now the Grok investigation arrives at a tense moment in
- 35:37transatlantic relations. US Secretary of State Marco
- 35:39Rubio responded to the December EU find by saying the block is
- 35:43attacking American companies, and a State Department official
- 35:46warned that nothing is off the table when it comes to free
- 35:48speech. The EU has made clear that its
- 35:51focus is changing platform behavior rather than simply
- 35:54collecting fines. The Digital Services Act gives
- 35:57regulators the power to impose operational requirements, demand
- 36:02specific changes to product features, leave it suspend
- 36:04services that repeatedly violate European law.
- 36:08No other X will face such drastic measures.
- 36:11Depends on the investigations, findings and the company's
- 36:13willingness to implement meaningful safeguards.
- 36:22New financial filings just revealed what Elon Musk probably
- 36:24didn't want you to see. X, formerly Twitter, of course,
- 36:29lost 58% of its revenue in 2024. Now documented an official UK
- 36:35regulatory filings that dropped this week that we spent the last
- 36:39two years hearing about advertisers fleeing, about the
- 36:42rebrand chaos, about whether Musk's social media gamble would
- 36:46pay off. Now we have the answer, and it's
- 36:48brutal. This is particularly fascinating
- 36:51because 2024 was supposed to be Musk's victory lap.
- 36:55He became the most politically influential tech billionaire in
- 36:58America, practically lived at Mar a Lago with Donald Trump,
- 37:03helped shape a presidential administration as well.
- 37:06His visibility had never been higher, and yet X was bleeding
- 37:11out. And I've also got some news on
- 37:14the EU launching a formal investigation into Grok over
- 37:17deep fake porn. Plus a genuinely surprising
- 37:20partnership between Tesla and Warren Buffett's company, and
- 37:24we'll go right into that after this very short break.
- 37:29So let's talk about what a 58% revenue decline actually means
- 37:33for X. When Musk bought Twitter in late
- 37:3520/22 for 44 billion, a lot of people thought he overpaid for
- 37:40it. Now, the bull case was that he'd
- 37:42unlock value by cutting costs, removing what he called bloat,
- 37:46and turning the platform into an everything app.
- 37:49And the bear case was that advertisers would flee, the
- 37:52platform would become toxic, and the whole thing would spiral out
- 37:56of control. Now, these UK filings suggest
- 38:01that the bears were right. Now, these are specifically UK
- 38:04revenue numbers, but they're a proxy for the broader health of
- 38:08the company because the UK is a major advertising market.
- 38:11And when brands have pulled back, they're pulling back kind
- 38:15of everywhere. It's a it's a litmus test.
- 38:19And here's what's interesting about this.
- 38:21Musk has been adamant publicly that X is doing fine.
- 38:25He's pointing to engagement numbers, to new features, to the
- 38:28premium subscription tier. But engagement revenue are
- 38:33completely different things. Get 1,000,000 of people
- 38:36scrolling and still watching your ad business collapse.
- 38:39Advertisers don't just want your eyes on the product, they want
- 38:44it to be safe and they want predictability.
- 38:47They want to not wake up to headlines about their logo
- 38:49appearing next to something that makes their PR team cry tears of
- 38:54sorrow. And X under Musk has been the
- 38:58opposite of predictable. The timing makes this even more
- 39:02striking. Throughout 2024, Musk was
- 39:04everywhere, posting constantly, engaging with political figures,
- 39:09hosting Twitter Spaces with presidential candidates.
- 39:12If personal brand translated directly to business success, X
- 39:16should be printing money. Instead, the opposite happened.
- 39:20The more visible Musk became, the more nervous advertisers
- 39:23got. There's a lesson in there
- 39:25somewhere about the difference between attention and trust.
- 39:29Now, I trust that you are subscribed to this channel.
- 39:33I've been digging through the analytics of the show and
- 39:35noticed that 39% of you are following the show and I am
- 39:38forever grateful for you. And the other 61% of you have
- 39:42probably listened to a few episodes, just forgot to hit the
- 39:45subscribe or follow button. So if you could do that right
- 39:47now, that'd be great. I've been an independent
- 39:49journalist covering Elon and Space Flight and tech for the
- 39:51last seven years. I'm going to continue for the
- 39:54next 10. And all I ask from you is one
- 39:56second of your time. I'm extremely grateful for you
- 39:59to be here. Thank you for joining this
- 40:01community. Now let's move on while X deals
- 40:06with its revenue crisis. Back in the day, another Musk
- 40:11venture just landed in regulatory crosshairs.
- 40:14The European Union announced a formal investigation into Grok,
- 40:18Musk's AI chatbot, through XAI over concerns of deep fate
- 40:22pornography. This isn't a sternly worded
- 40:25letter or a tweet from some EU commissioner.
- 40:30It's a formal probe under EU digital services regulations,
- 40:34and it carries real teeth. The issue is that Grok has been
- 40:38notably less restricted than competitors like ChatGPT or
- 40:41Claude Must position this as a feature, not a bug.
- 40:45He's been critical of what he calls overly censored AI,
- 40:48arguing that other companies are too cautious, too politically
- 40:51correct. Grok was supposed to be the edgy
- 40:54alternative. The problem is that edgy and
- 40:57generate sexualized deepfakes are different, and the EU isn't
- 41:01interested in debating philosophical approaches to AI
- 41:04freedom. They passed the AI Act, they've
- 41:06been the most aggressive regulator on the planet when it
- 41:09comes to tech, and now they're coming for Grok.
- 41:12This matters beyond just the regulatory fine that came.
- 41:16The EU is a massive market. If Grok gets restricted or
- 41:20banned in Europe, that's a significant chunk of potential
- 41:23users gone, and it creates a template for other regulators
- 41:26around the world. The UK might actually follow
- 41:29Australia. So making noise about AI
- 41:32regulation in the US probably not going to happen there.
- 41:35But once one major jurisdiction acts, others tend to pile on.
- 41:39So this isn't just a European thing.
- 41:42This is about whether Musk's approach to AI, which is
- 41:45basically move fast and let's figure this out eventually, can
- 41:49survive contact with actual governments.
- 41:52And what strikes me is the Pattern X is struggling because
- 41:56Elon prioritized his vision of free speech over advertiser
- 42:00comfort. He didn't think of the business
- 42:02model, he just did what he wanted to do.
- 42:04Grok is in trouble because Musk prioritized his vision of
- 42:07uncensored AI over regulatory expectations.
- 42:11There's a consistency to it. He genuinely believes these
- 42:14institutions, advertisers, regulators are all wrong.
- 42:17And he's right. Maybe he is, but being right and
- 42:21being financially viable aren't always the same thing.
- 42:24Now, let's kind of move into something more positive here
- 42:28because this week also brought news that Warren Buffett's
- 42:31company willpower test the semi trucks.
- 42:34Warren Buffett, the guy who famously avoids tech stocks, the
- 42:37guy who said he doesn't invest in things he doesn't understand.
- 42:40The guy who's been skeptical of Tesla, specifically that Warren
- 42:45Buffett's operation is now partnering with Elon Musk.
- 42:49How? The deal involves Berkshire
- 42:50Hathaway's energy subsidiary providing charging of structure
- 42:54for Tesla Semi deployments. This is significant for a few
- 42:58reasons. First, it suggests that Tesla
- 43:00Semi is finally moving towards serious commercial scale.
- 43:03The truck has been in development limbo for years,
- 43:06with production starts and stops questions about battery range.
- 43:09Skepticism from the trucking industry getting a Buffett
- 43:12company involved signals that serious old school money
- 43:17investors see a real business model here. 2nd, it's a
- 43:21validation of Tesla from someone who has historically been
- 43:24allergic to the company. Buffett and Musk occupy
- 43:27completely different corners of the business world.
- 43:30Buffett is patient capital, long time horizons, dozens and dozens
- 43:36of years in the making, boring industries.
- 43:39Musk is move fast, blow rockets up, sleep at the factory, do
- 43:44whatever you got to do to make this thing work.
- 43:46The fact that Buffett's team looked at Tesla Semi and said
- 43:49yes we want it tells you something about how the industry
- 43:51views that tech now. And 3rd, this is exactly the
- 43:55kind of institutional partnership Tesla needs as it
- 43:58tries to diversify beyond consumer vehicles.
- 44:00The Tesla Semi targets commercial fleets, logistics
- 44:04companies, enterprises with massive fuel budgets.
- 44:07Those customers aren't swayed by Twitter posts or cyber truck
- 44:10aesthetics. They want proven tech and
- 44:12reliable infrastructure. Having Berkshire Hathaway's
- 44:15energy arm involved checks both of those boxes.
- 44:19So here's where we are with the Musk empire as we close out
- 44:23January 2026. Today, X is hemorrhaging money.
- 44:26Back in the day, even as Musk's personal influence peaked, his
- 44:31AI company faces a formal EU investigation that could reshape
- 44:34how Grok operates in major markets.
- 44:36But Tesla? His oldest company just locked
- 44:40in a partnership with one of the most conservative, credibility
- 44:42focused investors on the planet. The empire is simultaneously
- 44:46expanding and contracting depending on where you look.
- 44:49Question for the rest of this year is whether Musk's political
- 44:52capital can translate to business stability or whether
- 44:56the financial cracks, enacts and regulatory pressure on Grok will
- 44:59eventually matter more than how many times that Elon posts in a
- 45:03day. 75 countries. That's how many nations sent
- 45:08representatives to 1 conference this week not to talk about
- 45:11nuclear weapons or climate change or trade deals.
- 45:14Talk about space junk. The Saudi Space Agency hosted
- 45:18the second annual space debris conference.
- 45:20If you want to sign. The commercial space industry is
- 45:22actually growing up. This is the time.
- 45:25The gold rush phase launch fast, grab your orbit, figure out the
- 45:29mess later is giving way to something more sobering.
- 45:32A global reckoning with what we've left behind up there in
- 45:35the companies building mega constellations like Starlink,
- 45:38They are the elephant in the room.
- 45:40So today I want to walk through these stories that, taken
- 45:43together, paint a picture of where Space Flight is right now.
- 45:47We've got the debris problem demanding attention.
- 45:49We've got NASA pushing ahead with Artemis 2 getting ready to
- 45:52send humans around the moon. And we've got nuclear propulsion
- 45:55tests that could reshape how we think about deep space travel
- 45:59entirely. Growth and accountability happen
- 46:02at the same time. And that's what a maturing
- 46:04industry looks like. That's what the Space Flight
- 46:06industry looks like right now. We're going to get into that
- 46:09right after this very short break.
- 46:13So let's start with this debris conference.
- 46:15The numbers here are massive. 75 countries showed up.
- 46:20The event was organized by the Saudi Space Agency, but it had
- 46:23backing from the United Nations Office for Outer Space Affairs
- 46:27in partnership with the International Telecommunication
- 46:30Union. Here's what Artie Holimany, the
- 46:35director of UNOOSA, said. Global cooperation among
- 46:39international organizations, governments, industry and
- 46:42academia in addressing space debris challenges is no longer
- 46:45merely A desirable objective is become an absolute necessity.
- 46:51She also said that leading space faring nations must develop
- 46:54concrete plans to address accumulated legacy debris.
- 46:59Legacy debris is a polite way of saying the stuff we've been
- 47:02leaving up there for decades. All the junk.
- 47:05Now why does all this matter though?
- 47:07The International Telecommunication Union manages
- 47:09satellite orbit allocations. They're the people who
- 47:13coordinate who gets to park where in space.
- 47:17When they show up at a debris conference, signals that orbit
- 47:20management and debris are now in the same conversation.
- 47:23They were kind of a side thing. That debris thing was for a
- 47:26while, but now it's all coming together.
- 47:29Can't talk about where new satellites go without talking
- 47:31about what's already cluttering up those lanes.
- 47:35Now. Space is huge.
- 47:37And here's where SpaceX comes into the picture, even though
- 47:41they weren't mentioned by name. Starlink has thousands of
- 47:44satellites in low Earth orbit. They want to get hundreds of
- 47:47thousands into low Earth orbit. Thousands more are planned for
- 47:50the next few years. Amazon's working on their own
- 47:53consolation. China is doing the same thing.
- 47:56It's getting uncomfortable up there really fast.
- 47:59And more objects mean more collision risk.
- 48:02If something happens to 1 satellite, if one SpaceX
- 48:05Starlink satellite breaks, a piece from that Starlink
- 48:10satellite could destroy numerous satellites.
- 48:12Could be a chain reaction. More collisions equal more
- 48:16debris, and more debris means more collision risk.
- 48:19It's a feedback loop that keeps people up at night From the
- 48:23UNOOSA. They also featured something
- 48:27called the Debris Solver Competition at that conference,
- 48:3120 projects showcasing innovative solutions for
- 48:34actually cleaning up space junk. We're not just wringing our
- 48:39hands with the problem, just saying like, hey, there's debris
- 48:42up there. Who cares?
- 48:43It's space. It's not down here.
- 48:45There's engineers and people that want to make money,
- 48:48entrepreneurs actively working on fixes for this.
- 48:53They want to make Nets, harpoons, there's lasers,
- 48:56robotic arms. If there's space junk, there's
- 49:00going to be money to be made. It's like having a garbage
- 49:04person pick up your garbage in front of your house.
- 49:06There's somebody there that does that.
- 49:08There's a business that makes a lot of money doing garbage
- 49:11pickup, and so these entrepreneurs want to do too.
- 49:14They want to pick up the garbage in space.
- 49:15They want to haul it someplace else.
- 49:18So some of these ideas sound crazy, right?
- 49:22Sound like science fiction, But so did Spacex's reusable rockets
- 49:26for about 10 years ago. Now, something else crazy is
- 49:29that I've been digging into the analytics of the show, and 37%
- 49:32of you are following the channel and I am forever grateful for
- 49:36you. The other 63% haven't hit the
- 49:39follower subscribe button. It's OK.
- 49:41Sometimes you listen to a show you're not subscribed.
- 49:43Sometimes you listen to a few episodes and it's cool, but
- 49:47eventually you do come around and hit the subscribe button.
- 49:49That's what I do all the time. I'd listen to a show four or
- 49:52five times and I'm like, oh wait, I should subscribe.
- 49:54So I don't have to search for this anymore.
- 49:55Just comes into my feed and that's going to help us out
- 49:58tremendously. So all I ask from you, after six
- 50:02or seven years of already doing this, I ask one second of your
- 50:05time. So thank you for that for
- 50:07hitting the subscribe or follow button.
- 50:09So there's a reckoning going on too.
- 50:13There's an execution with what NASA is actually doing right now
- 50:19about this. Artemis 2 is moving forward very
- 50:23soon. February, NASA announced they're
- 50:25pushing ahead with what's called a wet dress rehearsal for the
- 50:28mission. And a wet dress rehearsal is
- 50:31essentially a full simulation of launch day.
- 50:34You load the rocket with propellants, that's the wet
- 50:36part, Run through the entire countdown sequence and stop just
- 50:40before ignition. They do everything they need to
- 50:42do before launch, and then they just don't launch the rocket.
- 50:45It's how you find problems before they become catastrophic.
- 50:48If there's any leaks, they find them before the actual launch
- 50:51day. This is huge because Artemis 2
- 50:54is the mission that sends humans around the moon for the first
- 50:57time since Apollo 4 astronauts. A loop around the far side of
- 51:01the moon and the first crude flight of the Space Launch
- 51:04System and Orion capsule working together.
- 51:07And that's after Artemis 1 proved the hardware could make
- 51:10the trip on crude. This is the next step.
- 51:12And then Artemis 3 should have people stepping on the moon
- 51:17surface again. Now, there's also a Utah
- 51:20connection worth mentioning here.
- 51:22The solid rocket boosters, the power SLS during liftoff are
- 51:25manufactured there. Utah has been part of NASA's
- 51:29human spaceflight story since the space shuttle days.
- 51:32So we talked about Artemis. We're talking about supply
- 51:34chains and expertise that span across the country, and the
- 51:38University of North Dakota is also involved in supporting this
- 51:42mission. When you see university is
- 51:44embedded in NASA programs, it's something about the depth of
- 51:48those efforts. Together, the university and
- 51:50NASA working hand in hand, it's huge.
- 51:53They're not just building rockets, though.
- 51:55They're working all across the nation, pulling in academic
- 51:59research, industry partners in multiple states.
- 52:03So the Artemis 2 mission needs to get through the orbit that
- 52:09the debris fields are in now. They will plan that properly so
- 52:13nobody gets hurt, but the Artemis missions going forward
- 52:18will need to clean up some of this garbage.
- 52:22Artemis 2 may face delays if the wet dress rehearsal doesn't work
- 52:27out as planned. The heat shield issues from
- 52:29Artemis One took a lot longer to resolve than anybody wanted to
- 52:32talk about. In getting through this
- 52:34rehearsal successfully would be signal that the program is way
- 52:39back on track. So let's go beyond the moon
- 52:42here. NASA is testing advances in
- 52:45space nuclear propulsion. This all matters because it's
- 52:48all tied into the orbit of all the space debris.
- 52:53This is going to be a technology that could genuinely change the
- 52:58game for deep space exploration. Chemical rockets are great for
- 53:02getting off Earth. They produce tremendous thrust,
- 53:05but once you're out in space, they're inefficient for long
- 53:09journeys. You burn through fuel really
- 53:12fast, and you can only carry so much of it.
- 53:15And once that fuel is gone, the fuel is gone.
- 53:18You can't get more. Nuclear propulsion is a whole
- 53:23different beast. Nuclear thermal rockets heat
- 53:26propellant to extreme temperatures using a reactor,
- 53:30generating far more efficiency per unit of fuel.
- 53:34And the result of that is that you could cut the transit time
- 53:36to Mars significantly. Some estimates suggest getting
- 53:40there in three to four months instead of 7 to 9.
- 53:44That's huge for safety. Less time in transit means less
- 53:47radiation exposure for astronauts.
- 53:50Fewer things can go wrong if you shorten the time frame.
- 53:54And NASA has been researching this for years.
- 53:56But recent tests suggest real progress there and it's not
- 54:01going to be ready next year, but we might be talking about it,
- 54:05getting ready for a Mars mission that are currently on the
- 54:09drawing board sometime in the 20 thirties.
- 54:12So this is the context here. We're serious about becoming a
- 54:14multi planet spirit planetary species, which is something both
- 54:18NASA and SpaceX talk about constantly.
- 54:21We need propulsion systems that can handle the distances
- 54:24involved. Chemical rockets can get us to
- 54:26the moon. Nuclear propulsion might be what
- 54:29gets us to Mars and beyond any practical time frame.
- 54:34Now we have to talk about one more thing.
- 54:36Rocket Lab, a small company that's sort of pseudo competitor
- 54:42to SpaceX, their Neutron rocket experienced a test failure
- 54:47recently. We don't have all the details
- 54:49yet, but analysis seems to be maintaining confidence in the
- 54:53company despite this set back because they're going to fix it.
- 54:57It's notable because Rocket Lab is one of the few companies
- 54:59seriously competing against SpaceX in the medium lift
- 55:03market. Neutron is their answer to
- 55:05Falcon 9. But, you know, test failures.
- 55:09They're part of the rocket development programs.
- 55:11SpaceX blew up plenty of prototypes on their way to
- 55:14success. And now look at the Falcon 9.
- 55:16Hundreds of launches, hundreds of landings.
- 55:18The question is always whether a company can learn from the
- 55:21failure and move on. And Rocket Lab has had a strong
- 55:24track record with their smaller Electron rocket, so I'm not
- 55:27going to count them out. But this business is hard, and
- 55:31they will learn from their mistakes and they'll move on
- 55:34from it. So where does this all tie
- 55:36together? The same week that 75 countries
- 55:39gathered to talk about cleaning up Earth orbit, NASA was testing
- 55:43the nuclear engines and preparing to send humans back to
- 55:46the moon. This is progress, and it's going
- 55:50to be a little bit longer, but we're going to get there.
- 55:53And they're simultaneously pushing boundaries and dealing
- 55:56with consequences the same exact time.
- 56:00The early years of any industry are about proving the tech that
- 56:03actually works. And now the middle years are
- 56:06scaling it, and we're heading into the later years of this.
- 56:13So they need to sustain it. Commercial spaceflight is
- 56:16entering that third phase. The launches will keep coming.
- 56:20Hundreds of SpaceX Starlink flights will happen every year.
- 56:24Starship will fly and it'll take thousands of Starlinks into
- 56:30orbit, amongst other things. Starlink will grow, but so will
- 56:33the other competitors, and so will the conversations about
- 56:36responsibility. Question for SpaceX and every
- 56:39other player in the industry is whether they can keep doing this
- 56:44and keep launching without somebody there to clean up their
- 56:47mess. And who's going to pay those
- 56:49people in those jobs to clean up this mess?
- 56:52We're going to have to find that out in the future.
- 56:54Is it going to be the country's or will the will the companies
- 56:59be taxed for clean up over the time that they have their
- 57:03satellites in orbit? Tik Tok's algorithm broke over
- 57:08the weekend. Users open the app Sunday and
- 57:10found their hyper personalized feeds replaced by generic videos
- 57:15on repeat. Now the company says a power
- 57:17outage at AUS data center triggered A cascading systems
- 57:21failure. Creators watching their view
- 57:23counts dropped to 0, earnings vanishing from all of their
- 57:27dashboards. The timing here is critical.
- 57:31This happened just days after TikTok finalized this deal to
- 57:34spin off its U.S. business into a new entity controlled by
- 57:37American investors and more importantly, Trump supporting US
- 57:44American investors. So is this a technical glitch or
- 57:47is this something bigger that we should be really paying
- 57:50attention to? Now we're going to break down
- 57:53exactly what happened to that data center, why the algorithm
- 57:56failing matters more than a typical outage, and what this
- 57:59tells us about the new TikTok. USDS joint venture also going to
- 58:04look at a public fight between Elon Musk and billionaire
- 58:07investor Vinod Khosla over accusations of white
- 58:11nationalism. Plus, new data showing Tessa's
- 58:14brand value dropped in 2025 because of Musk's politics.
- 58:19Now, that part connects directly to the TikTok story in ways you
- 58:22might not expect, and we'll go right into that after this very
- 58:27short break. So let's start with what
- 58:30actually happened to TikTok on Sunday.
- 58:33A power outage hit one of the company's U.S. data centers.
- 58:36The company confirmed the outage caused a cascading systems
- 58:39failure. Now that means one system going
- 58:42down triggered failures in connected systems like dominoes,
- 58:45spreading the damage across the apps infrastructure.
- 58:48If you've ever seen a circuit breaker trip and knockout power
- 58:51to half of your house, this is the software equivalent, but
- 58:55affecting millions of people, not just you and your family.
- 58:58And the problem showed up immediately in ways users could
- 59:01feel. People reported trouble logging
- 59:04in. Video uploads failed.
- 59:06But the stranger system from the algorithm itself, Tik Tok's
- 59:10recommendation engine, the thing that makes the app feel like it
- 59:13actually knows you, started serving random, generic content.
- 59:17Some users saw the same few videos repeating in their feeds
- 59:19over and over. For an app that built its entire
- 59:22business on hyper personalized recommendations, this is a core
- 59:26failure. It changes how users experience
- 59:29the product at the most fundamental level.
- 59:32At the foundation, creators got hit the hardest.
- 59:36New uploads showed 0 views and 0 likes.
- 59:39Earnings dashboards went blank. People thought they lost
- 59:43hundreds of thousands of dollars.
- 59:45Tiktok said this was a display error caused by server timeouts,
- 59:49and the actual engagement data remains safe.
- 59:51But for creators who depend on real time metrics to understand
- 59:54their audience and make changes on the fly, several days of
- 59:57broken data feels like flying completely blind.
- 1:00:02Now. The time he makes this more than
- 1:00:04just a technical story, of course.
- 1:00:06TikTok finalizes deal to spin off the US business just days
- 1:00:09before the outage. The new entity, TikTok USDS
- 1:00:12joint venture, will be controlled by US investors, most
- 1:00:16of them Trump supporters. Users had already noticed the
- 1:00:19company pushing new terms of service and new privacy policies
- 1:00:22immediately after the deal closed.
- 1:00:24And then then the algorithm broke.
- 1:00:26People are connecting the dots, whether or not those dots
- 1:00:29actually connect, and then raises questions the company you
- 1:00:33will need to answer, even if the outage was entirely unrelated to
- 1:00:37ownership. Now, the US joint venture has
- 1:00:41said it plans to retrain Tik Tok's recommendation algorithm.
- 1:00:45That algorithm is the actual product of TikTok.
- 1:00:48There's nothing better than the TikTok algorithm on the market
- 1:00:52right now. And if you have to retrain that,
- 1:00:54that's not actual TikTok. That's something else.
- 1:00:57It is what makes TikTok different than every other video
- 1:00:59platform. If the current outage is giving
- 1:01:02us a preview of what a broken or retrained algorithm looks like,
- 1:01:06we're not going to like it. Generic videos on repeat is not
- 1:01:09the experience that made TikTok the most downloaded app in the
- 1:01:12world for years. Company said Monday had made
- 1:01:15significant progress on recovery, but warned US users
- 1:01:18might still have technical issues posting the videos.
- 1:01:22You know, I've been digging through the analytics of the
- 1:01:23show and I've seen 39% of you have followed the channel or
- 1:01:28liked it. For you, I'm forever grateful.
- 1:01:30And the 61% of you have listened to an episode before or have
- 1:01:36not, and you're new here, but you haven't hit the follower
- 1:01:39subscribe button. I get it.
- 1:01:40I do it all the time. If you're not going to hit the
- 1:01:43follower subscribe button after the first one, I've done that
- 1:01:46too. Sometimes it takes 3 or 4
- 1:01:48listens, four or five episodes. But I'm just asking you for a
- 1:01:51favor. I've been doing this for seven
- 1:01:53years. I'm going to be doing it for the
- 1:01:54next 10. I ask you for one second of your
- 1:01:56time, hit the follow button. If you want to leave a five star
- 1:02:00rating on whatever app you're on right now.
- 1:02:03I'm extremely grateful and blessed to have you in this
- 1:02:05community. Now let's get back into some
- 1:02:07Elon news and let's shift here. Completely different story.
- 1:02:11They happened over the same weekend.
- 1:02:13Elon got into a public fight with Vinod Khosla.
- 1:02:16He's a billionaire venture capitalist, Co founder of Sun
- 1:02:19Microsystems and one of the most influential investors in Silicon
- 1:02:22Valley. He accused Musk of promoting
- 1:02:25white nationalism. Musk fired back on X with a
- 1:02:28personal response. He said his partner's half
- 1:02:31Indian. He said his son's middle name is
- 1:02:33Sarkar, that is a name with Indian origins.
- 1:02:37Accusations of racism or white nationalism against ACEO can
- 1:02:40affect the companies they lead. It's not abstract, happens all
- 1:02:44the time. You have data showing it is
- 1:02:46already happening to Elon. A recent report found Tesla's
- 1:02:50brand value dropped significantly in 2025 due to
- 1:02:54Musk's political activities. Brand value measures how much
- 1:02:57the company name itself contributes to sales and
- 1:03:01customer loyalty. Like Tesla.
- 1:03:03We all know what that is. That's very, you know, they're
- 1:03:06very powerful and loyal. People are going to continue
- 1:03:10buying Teslas like Subaru. For me, I, it's a great car.
- 1:03:13I love Subarus. I want a Subaru next time.
- 1:03:16You know, when that number falls down, though, Subaru's CEO was
- 1:03:21called a white nationalist. I'd be like, probably going to
- 1:03:23look someplace else. It puts pressure on everything
- 1:03:26from pricing to employee recruitment.
- 1:03:29I would not work for somebody who's a white nationalist.
- 1:03:31That's a scummy kind of person, right?
- 1:03:33I don't want to work for that guy, but let's see.
- 1:03:37Let's see what's up with Musk. Khosla Exchange is showing that
- 1:03:41Musk's public persona has become inseparable from his brands.
- 1:03:46Now it's not just an EV company, you know, it's not just Tesla
- 1:03:51and Elon. It is Elon.
- 1:03:56SpaceX is Elon, X is Elon. XAI is Elon.
- 1:04:02When Musk gets into political fights, those fights follow him
- 1:04:06around and follow his companies around limits how the companies
- 1:04:10can position themselves to different consumer groups.
- 1:04:14The TikTok story in the Musk story share something important.
- 1:04:18Both show how leadership decisions ripple outward into
- 1:04:21user experience and brand perception.
- 1:04:23Tik Tok's new ownership structure is forcing technical
- 1:04:26changes that users can feel directly in their feeds.
- 1:04:29Musk's political choices are forcing brand changes that show
- 1:04:32up directly in the data. Neither situation is purely
- 1:04:36technical or purely political. They mix together in ways that
- 1:04:40affect real people, making real decisions about which real
- 1:04:44products that they use. And the bigger pattern is about
- 1:04:49controlling consequences. Tik Tok's US users are
- 1:04:52discovering what it feels like when the infrastructure behind
- 1:04:55their favorite app belongs to someone new with different
- 1:04:58priorities. Tesla customers are discovering
- 1:05:01what it feels like when the CEO behind their favorite car has
- 1:05:05political positions they disagree with.
- 1:05:07And in both cases the product experience changed because
- 1:05:10something changed at the top.