ChatGPT and Beyond with Fexingo: Large Language Models, Generative AI, and Productivity Tools

Episode 54 of ChatGPT and Beyond with Fexingo explores the quiet but powerful shift toward synthetic data in AI training. Lucas and Luna break down why companies like Microsoft and Meta are generating artificial training examples, how this reduces costs and privacy risks, and what happens when models start eating their own tail — a phenomenon known as model collapse. With real numbers on data cost...

SpaceX went public this week and raised $85.7 billion in its IPO, dwarfing every other IPO in history. Lucas and Luna break down how the deal worked, why the company chose a direct listing over a traditional IPO, and what that massive cash pile means for the space economy. They also look at what the market reaction tells us about investor appetite for high-risk, long-duration assets in mid-2026, w...

Episode 52 of ChatGPT and Beyond with Fexingo examines a growing concern in AI: model collapse, where AI systems trained on AI-generated data degrade and lose diversity. Lucas and Luna discuss how this phenomenon threatens long-term model performance, using real examples like NVIDIA's recent GPU sales and the rise of AI-generated content on the web. They explore why companies like Microsoft and Me...

This week, the IPO pipeline is packed with AI companies rushing to go public before sentiment shifts further. Lucas and Luna break down the numbers: over $12 billion in AI IPOs waiting to price, from data-center operator Switch to inference chip startup MatX. They discuss how the recent sell-off in software and infrastructure stocks—Adobe down 17%, Oracle down 13%, Super Micro down 31%—is creating...

In this episode, Lucas and Luna explore a surprising career trend: the rise of AI auditors. With KPMG recently pulling a report due to apparent hallucinations, and regulators circling big AI labs, companies are scrambling for people who can audit models for bias, accuracy, and compliance. Lucas explains what an AI auditor actually does, why the role is growing, and how much they can earn. He cites...

A KPMG report pulled over hallucinations last week. OpenAI is under state investigation. As AI deployment surges, a new professional is emerging: the AI auditor. Lucas and Luna break down why model evaluation, bias testing, and regulatory compliance are becoming a $12 billion market by 2028, and why companies like Microsoft and Palantir are scrambling to certify their systems. They also look at th...

Super Micro Computer's stock plunged over 30% in five days, far worse than any other AI hardware name. In this episode, Lucas and Luna dissect what's behind the drop — from the delayed 10-K filing and Hindenburg short report hangover to the broader rotation away from server makers toward ASIC designers like Broadcom. They examine SMCI's margin profile, the impact of the NVL72 rack timing, and what...

Lucas and Luna dig into the brutal June 2026 sell-off in AI software stocks, using Adobe, Salesforce, and ServiceNow as case studies. With Adobe down 16% in a week despite a strong quarter, they explore what's really driving the divergence between earnings performance and market sentiment. The conversation covers rising expectations, inference cost commoditization, and the shift from software lice...

Lucas and Luna explore the fast-growing technique of model merging in AI — combining the weights of multiple fine-tuned models into one without retraining. They discuss how Sakana AI's 'Evolutionary Model Merge' method won top spots on the LMSYS leaderboard with zero GPU training hours, why startups like MergeKit and Arcee AI are attracting VC interest, and why enterprises should care: better perf...

In this episode, Lucas and Luna explore the under-the-radar technique of AI model distillation—where a smaller, cheaper student model is trained to mimic a giant teacher model like GPT-5 or Claude 4. With enterprise AI spend hitting $7,500 per employee per month at some firms, distillation offers a path to slash inference costs by 10x while retaining 95% of accuracy. Using live data showing Super...

Episode 44 of ChatGPT and Beyond with Fexingo — Lucas and Luna explore the emerging technique of model merging, where developers combine two or more open-source neural networks without additional fine-tuning. They use a concrete example: merging a base Llama model with a specialized math model yields a model that scores 15 percent higher on GSM8K benchmarks than either parent. The hosts discuss th...

A new TechCrunch report reveals that 'AI-pilled' firms are spending an eye-popping $7,500 per employee per month on AI tools. Lucas and Luna break down where that money goes — inference compute, API subscriptions, fine-tuning — and why it signals a permanent shift in enterprise budgeting. They also examine the selloff in AI stocks like AMD, down 12.9% in five days, and SMCI, cratering 37.3%, quest...

Lucas and Luna explore the rapid shift of AI inference from cloud data centers to edge devices and on-prem servers. They discuss why companies like Apple and Meta are investing in local AI processing, the cost savings of reducing cloud dependency, and how this trend could reshape the hardware market. Specific examples include Apple's on-device AI features from WWDC 2026 and Meta's new data center...

In this episode, Lucas and Luna dive into the surprising trend of AI inference costs falling even faster than training costs. They explore the implications for enterprise adoption, the shift from GPU to custom silicon, and what it means for the AI market in mid-2026. With NVIDIA down 3.7% and AMD down 13.5% in the past week, they question whether the hardware narrative is changing. Plus, they disc...

Lucas and Luna explore the rising importance of AI model evaluation — a market that barely existed two years ago but is now attracting serious venture capital and enterprise spend. They discuss how companies like Scale AI and newcomer Arize AI are capitalizing on the need to benchmark, test, and monitor large language models in production. With the backdrop of a brutal week for AI infrastructure s...

Apple just wrapped WWDC 2026 and announced a wave of on-device AI features that could reshape how millions interact with artificial intelligence. Lucas and Luna break down the three biggest announcements: Siri AI that can finally understand context, the revamped Image Playground that doesn’t produce weird results anymore, and the new AI-powered Shortcuts workflows. They discuss why Apple's approac...

The AI trade is suddenly reversing. In this episode, Lucas and Luna dissect the brutal sell-off sweeping AI infrastructure and software names in early June 2026. With Broadcom down 16% in a week, ServiceNow off 17%, and Snowflake falling 15%, they trace the trigger to a single earnings miss from a bellwether enterprise AI player. Lucas explains why investors are recalibrating their expectations fo...

Lucas and Luna dig into the sharp sell-off hitting AI stocks in early June 2026, with the Nasdaq down over 10% in a week and names like ServiceNow, Broadcom, and Palantir dropping 15-17%. They explore two competing explanations: the valuation reset narrative versus the AI spending fatigue narrative. Lucas brings up the recent JOLTS data showing a cooling labor market and rising unemployment claims...

Lucas and Luna dive into OpenAI's newly announced Lockdown Mode, a feature designed to protect sensitive data from prompt injection attacks. They explore how this security measure works, why it's necessary, and what it means for enterprise AI adoption. Drawing on recent market sell-offs in AI and software stocks, they discuss whether these security concerns are contributing to investor unease. The...

Lucas and Luna dive into the brutal sell-off hitting enterprise AI software stocks in early June 2026. With ServiceNow down 17%, Broadcom off 16%, and Palantir dropping 15% in just five days, the hosts unpack what's really driving the rotation. They examine the divergence between AI infrastructure stocks like NVIDIA — down only 8% — and the software layer that's been hit far harder. The episode fo...

The AI software sector is in turmoil. Over the past week, stocks like ServiceNow, Palantir, and Snowflake have dropped 15-17%, while Adobe and Microsoft fell 8-10%. Yet earnings reports have been solid. How can this be? In this episode, Lucas and Luna break down the disconnect: a rotation out of high-beta AI names into defensives, profit-taking ahead of a Fed decision, and the growing fear that AI...

Lucas and Luna dig into Supabase's stunning valuation jump from $5B to $10B in just eight months — a rare bright spot in a week when enterprise software stocks like ServiceNow, Salesforce, and Snowflake got crushed. They unpack what Supabase is (an open-source Firebase alternative), why developers are flocking to it, and whether the valuation tells us more about AI's hunger for data infrastructure...

In this episode of ChatGPT and Beyond with Fexingo, Lucas and Luna dig into a surprising limitation of today's large language models: they still can't reliably answer simple calendar or date-arithmetic questions. Using a concrete example from June 2026—asking a model what day of the week June 5 falls on—they explore why temporal reasoning breaks down for transformer-based architectures, even as th...

Lucas and Luna discuss the surprising shift in AI infrastructure construction, where tech giants like Meta are adopting cost-effective, rapid-deployment methods — including tent-like structures — to house their GPU clusters. With Meta building data centers in tents, NVIDIA's stock up 4% in a week, and ARM surging 10.8%, they explore what this means for the chip supply chain, hyperscaler competitio...

Every AI stock story right now is about NVIDIA versus AMD. But this episode argues the real action in June 2026 is happening in infrastructure-adjacent names that own the data-center plumbing. We start with Broadcom's 12 percent weekly surge and Oracle's 13 percent jump after a multiyear Google Cloud deal was reported for Lovable, an AI startup. Then we unpack how custom chip design (ASICs) and en...