Tech Disruptions

GPT-5.4: The Intern Who Can Run a Whole Department

March 13, 202619:01Tech Disruptions

This episode introduces OpenAI's new GPT-5.4 model, highlighting its transition from a sophisticated chatbot to a "digital colleague" capable of "agentic workflows." Listeners will learn that this AI can control a computer's mouse and keyboard, navigate desktop environments, and operate software, even outperforming humans on specific operational tasks. The discussion also covers its staggering 1-million-token context window, enabling it to process vast amounts of information.

Key Takeaways

Detailed Report

GPT-5.4 represents a fundamental shift in artificial intelligence, moving beyond sophisticated chatbots to become a "digital colleague" capable of autonomous action and complex workflow execution. This new model from OpenAI is poised to redefine how businesses operate and how knowledge work is performed.

Agentic Workflows and Native Computer Control

The core innovation behind GPT-5.4 is its focus on "agentic workflows." Unlike previous models that might specialize in coding or conversation, GPT-5.4 is a unified, general-purpose model that combines advanced reasoning, coding, and native computer use. This means it can understand a goal, break it down into steps, and then execute those steps by interacting directly with its environment and using software.

Crucially, GPT-5.4 is OpenAI's first general-purpose model with the built-in ability to control a computer's mouse and keyboard. It can navigate desktop environments and operate software, performing actions rather than just simulating them. On the OSWorld benchmark for desktop navigation tasks, GPT-5.4 achieved a 75% success rate, surpassing its predecessor GPT-5.2 (47.3%) and even the human average of 72.4%. This capability moves AI beyond brittle Robotic Process Automation (RPA) tools, offering robust, intelligent action.

A Million-Token Context Window

Another staggering advancement is GPT-5.4's 1-million-token context window. This is the AI's working memory, allowing it to process the equivalent of thousands of pages of documents in a single interaction. For context, early large language models (LLMs) had only a few thousand tokens. This massive increase enables the AI to ingest and analyze entire codebases, years of financial reports, or hundreds of research papers without losing context.

This expanded memory fundamentally changes the type of tasks AI can handle. Instead of breaking down large reports into small chunks for processing, GPT-5.4 can analyze entire projects, identify trends, correlations, and anomalies, and then generate comprehensive reports or presentations. This capability allows for the automation of entire multi-step processes, significantly streamlining knowledge work workflows.

The "Lost in the Middle" Problem

Despite the impressive context window, a known challenge with very long documents is the "lost in the middle" problem. Models can sometimes struggle to recall information if it's buried in the middle of a lengthy text, performing best with information at the beginning and end. While a million tokens is powerful, if critical information is missed, it poses a significant risk for high-stakes analysis. OpenAI has not explicitly detailed how GPT-5.4 mitigates this, suggesting human awareness and oversight remain crucial.

Enhanced Reasoning and Reliability

GPT-5.4 introduces a new "Thinking" or "extreme reasoning" mode, allowing the model to dedicate significantly more computational resources to solve complex, multi-step problems. Complementing this is "Upfront Planning," where the model displays its reasoning plan *before* execution. This transparency allows users to understand the AI's approach, make corrections, and steer its output, addressing the "black box" problem common in earlier LLMs and building trust for enterprise adoption.

Furthermore, OpenAI claims significant improvements in reliability, reporting a 33% reduction in factual errors and an 18% reduction in overall error rate compared to GPT-5.2. This addresses the notorious "hallucination problem," a major barrier for deploying AI in mission-critical fields like legal, finance, or consulting, where accuracy is paramount.

Strategic Implications: AI as a C-Suite Concern

With AI agents now capable of executing complex, multi-step business workflows from end to end, artificial intelligence is no longer solely an IT department concern; it's a strategic imperative for the C-suite. Deploying these "digital workers" requires a complete rethinking of business processes, organizational design, and how work is structured.

Major tech players like Microsoft (Copilot Studio), Google Cloud (Vertex AI Agent Builder), Salesforce (Agentforce), ServiceNow, and UiPath are all building the necessary infrastructure to deploy, manage, and govern these AI agents at scale. Their focus is on governance, safety, deep system integrations, and observability, recognizing the complexity of having thousands of AI agents making decisions and taking actions within corporate systems.

The Human-in-the-Loop Imperative

As AI capabilities grow, the "human-in-the-loop" (HITL) principle is becoming foundational for responsible AI deployment. While agents can execute routine steps independently, they are designed to pause and request human approval when encountering uncertainty or high-impact decisions. This approach blends the scalability and speed of AI with human nuance, emotional intelligence, and contextual understanding.

HITL also establishes clear lines of accountability, ensuring that a human is ultimately responsible for critical actions taken by an AI system. The goal is a symbiotic relationship where AI handles heavy lifting and flags issues, while humans provide judgment and oversight, rather than aiming for full automation.

Disruption in Knowledge Work Industries

The advent of highly capable AI agents will profoundly disrupt industries built on repeatable knowledge work and structured processes. Consulting firms, for instance, will see many tasks traditionally performed by junior consultants—such as information gathering, market research, data analysis, and report drafting—become highly automatable. This will likely lead to smaller, more productive consulting teams, with human roles elevating to higher-level strategic functions.

Business Process Outsourcing (BPO) is another industry facing massive transformation. Historically built on labor arbitrage, BPO will shift towards technology and intelligence, with AI agents automating data entry, invoice processing, and routine customer service. This will give rise to "AI-first" BPOs whose core value proposition is intelligent automation, offering 24/7 service and data-driven insights.

This shift also means a massive change in required skills for the workforce. Roles focused on routine, codifiable tasks are at high risk of automation. Conversely, roles demanding complex problem-solving, strategic thinking, creativity, and interpersonal skills will be augmented and become more valuable. The future of knowledge work will involve humans and their AI "digital colleagues" collaborating, necessitating significant re-skilling and transformation of the workforce.

Show Notes

GPT-5.4: The Intern Who Can Run a Whole Department

Source Materials

  • Research prompt on the capabilities and implications of a hypothetical advanced AI model, GPT-5.4, particularly its ability to control computers, handle large contexts, and its impact on agentic workflows and industries.

References & Resources

  • OpenAI: The company behind the GPT series of AI models, discussed as the developer of GPT-5.4.
  • GPT-5.4: The new, hypothetical advanced AI model discussed in the episode, featuring native computer control, a 1-million-token context window, "Thinking" mode, and reduced hallucinations.
  • GPT-5.2: The predecessor model to GPT-5.4, used for performance comparisons.
  • Google: A major tech company mentioned as having rival AI models.
  • Anthropic: Another major tech company mentioned as having rival AI models.
  • OSWorld: A benchmark specifically designed to evaluate AI models' ability to navigate and operate within desktop environments.
  • Microsoft: A major tech player building infrastructure for AI agents.
  • Copilot Studio: Microsoft's platform designed for integrating and managing AI agents within its ecosystem.
  • Microsoft 365: Microsoft's suite of productivity applications, part of the ecosystem where AI agents are being integrated.
  • Teams: Microsoft's communication and collaboration platform, part of the ecosystem where AI agents are being integrated.
  • Azure: Microsoft's cloud computing platform, providing the foundation for AI agent deployment.
  • Google Cloud: Google's suite of cloud computing services, offering platforms for AI agents.
  • Vertex AI Agent Builder: Google Cloud's platform for building and deploying AI agents.
  • Salesforce: A leading customer relationship management (CRM) platform provider, integrating AI agents into its services.
  • Agentforce: Salesforce's platform for AI agents focused on sales and customer service.
  • ServiceNow: An established enterprise software company integrating AI agents into its IT Service Management (ITSM) and other platforms.
  • UiPath: A leading Robotic Process Automation (RPA) platform provider, integrating AI agents into its automation solutions.
  • Consulting firms: An industry built on knowledge work, facing significant disruption and transformation due to advanced AI agents.
  • Business Process Outsourcing (BPO): An industry focused on outsourcing business functions, undergoing a shift from labor arbitrage to AI-driven services.
  • AI-first BPOs: A new type of BPO company whose core value proposition is intelligent automation and AI capabilities, offering autonomous routine task handling and data-driven insights.

Glossary

  • AI (Artificial Intelligence): The simulation of human intelligence processes by machines, especially computer systems. These processes include learning, reasoning, and self-correction.
  • LLM (Large Language Model): A type of artificial intelligence program that can recognize, summarize, translate, predict, and generate content using very large datasets.
  • Chatbot: A computer program designed to simulate conversation with human users, especially over the internet.
  • Agentic workflows: AI systems that can autonomously understand a high-level goal, break it down into a series of actionable steps, and then execute those steps by interacting with its environment and using various tools.
  • Context window: The amount of information (measured in "tokens") that an AI model can "remember" and process at one time during a single interaction or prompt. It's like the AI's working memory.
  • Tokens: The basic units of text that an AI model processes. These can be words, parts of words, or even individual characters, depending on the model's tokenizer.
  • Robotic Process Automation (RPA): Software technology that uses "software robots" to automate repetitive, rule-based tasks by mimicking human interactions with digital systems, often by interacting with user interfaces.
  • OSWorld: A specific benchmark or test suite designed to evaluate how well AI models can navigate and operate within a desktop operating system environment, performing tasks like opening applications or manipulating files.
  • Thinking mode / Extreme reasoning mode: A feature in advanced AI models that allows them to dedicate significantly more computational resources and time to solve complex, multi-step problems, often by exploring multiple reasoning paths.
  • Upfront Planning: A capability where an AI model explicitly outlines its intended steps or reasoning process *before* executing a task, allowing users to review, correct, or steer its approach.
  • Black box problem: The challenge of understanding *how* an AI model arrives at its decisions or outputs, making it difficult to interpret its internal workings, debug errors, or build trust.
  • Hallucination problem: A phenomenon where AI models confidently generate information that is false, nonsensical, or factually incorrect, often without any basis in their training data or input.
  • Lost in the middle problem: A known limitation in large context window LLMs where their performance in recalling or utilizing crucial information tends to degrade if that information is located in the middle of a very long input document, rather than at the beginning or end.
  • Serial position effect: A psychological phenomenon observed in humans where items presented at the beginning (primacy effect) and end (recency effect) of a list are remembered more accurately than items in the middle.
  • C-suite: Refers to the collective group of a company's most senior executive officers, such as the Chief Executive Officer (CEO), Chief Financial Officer (CFO), and Chief Operating Officer (COO).
  • Human-in-the-loop (HITL): An approach to AI deployment where human oversight and intervention are intentionally integrated into the AI workflow, especially for critical decisions, when the AI encounters uncertainty, or for validation.
  • Competitive moat: A sustainable competitive advantage that makes it difficult for rivals to compete with a business, protecting its long-term profits and market share.
  • Labor arbitrage: The practice of taking advantage of differences in wage rates between countries or regions by moving business operations or services to locations with lower labor costs.
  • AI-first BPOs: Business Process Outsourcing companies whose core business model and value proposition are built around leveraging advanced AI and intelligent automation, rather than primarily relying on low-cost human labor.

Full Transcript

HostOkay, so you know how we’ve been talking for years about AI becoming more capable, moving past just being a fancy search engine or a glorified chatbot?
ExpertOh yeah, the whole "AI is coming for our jobs" conversation, but mostly it's been about automating rote tasks or generating text. Interesting, but not exactly running a department.
HostExactly. But what if I told you there's a new model out, GPT-5.4, that doesn't just *talk* about doing work, it *does* it? And not just simple stuff, I mean it can literally control your computer's mouse and keyboard, navigate desktop environments, and operate software. It's already outperforming humans on specific software operation tasks.
ExpertWait, outperforming *humans*? Like, actual people trying to do the same task? That's… that's a pretty big leap from just writing a marketing email. It sounds less like an assistant and more like a digital intern who can actually, you know, intern.
HostMore like an intern who can run the whole department while you're on vacation. This isn't just an incremental upgrade, this is being positioned as a fundamental shift. We're talking about AI transitioning from a tool to a "digital colleague," capable of executing complex, multi-step workflows across various software applications.
ExpertWow. Okay, you've got my attention. "Digital colleague" is a loaded term, but "outperforming humans" on actual desktop tasks is something else entirely. What's powering this? What changed?
HostSo, let's dive in. The core idea here is that OpenAI, with GPT-5.4, has explicitly moved beyond just building really smart chatbots. They're focused on what they call "agentic workflows." Think of it this way: previous models might specialize in coding or conversation, but this one is a unified, general-purpose model combining advanced reasoning, coding, *and* native computer use. It's not just discussing work, it's actively performing it.
Expert"Agentic workflows" – that's the buzzword for this era, right? It means the AI can understand a goal, break it down into steps, and then execute those steps by interacting with its environment, using tools. So, GPT-5.4 isn't just generating text; it's *taking action* within software. That's a fundamental difference. It's the difference between telling a chef how to cook and the chef actually cooking.
HostPrecisely. And the specs behind this are pretty staggering. Let's start with the big one: a 1-million-token context window.
ExpertA *million* tokens? My jaw just hit the floor. For our listeners who might not be deep in the weeds on this, the context window is essentially the AI's working memory, right? How much information it can "remember" and process at one time. Early LLMs had a few thousand tokens, maybe enough for a couple of pages. A million tokens… that's like, a small library.
HostExactly! We're talking the equivalent of thousands of pages of documents. Imagine giving an AI your entire codebase, or an entire year's worth of financial reports, or hundreds of research papers – all in a single interaction. It can ingest and process all of that without losing its mind. And apparently, this matches what some of the rival models from Google and Anthropic are doing too, so it's becoming the new high-water mark.
ExpertThat's a game-changer for anything requiring deep analysis or synthesis across vast amounts of information. No more breaking things into tiny chunks and losing context. But the headline feature for me is still this native computer control. You mentioned it surpasses humans on benchmarks. How does that even work?
HostThis is where it gets wild. GPT-5.4 is OpenAI's first general-purpose model with the built-in ability to literally control a computer's mouse and keyboard. It can navigate desktop environments and operate software. Think about that for a second. It's not just simulating an action; it's performing it.
ExpertSo it's like a super advanced robotic process automation tool, but with actual intelligence behind it, not just pre-programmed macros. That's a huge distinction. RPA has been around, but it's very brittle; it breaks if the UI changes even slightly. This sounds much more robust.
HostIt is. The benchmark they use is called OSWorld, for desktop navigation tasks. GPT-5.4 achieved a 75% success rate. Its predecessor, GPT-5.2, was at 47.3%. And get this: the human average on that same benchmark is 72.4%. So, yes, it’s already slightly outperforming humans in specific software operation tasks. This isn't "interesting demo" anymore; this is "outperforming human capability."
ExpertThat's the "holy crap" moment right there. For years, we've talked about AI augmenting humans, doing the grunt work *for* us. But an AI that can navigate a desktop and operate software better than a human, even on specific tasks… that means the scope of "grunt work" just expanded exponentially. What else is under the hood?
HostThere's also a new "Thinking" mode, or "extreme reasoning" mode. This allows the model to dedicate significantly more computational resources to solve complex, multi-step problems. And critically, it has "Upfront Planning," where it displays its reasoning plan *before* execution. So, you can see how it's going to tackle a problem, make corrections, and steer its output without having to start all over again. That's a huge step towards usability and control for users.
ExpertThat's brilliant. That addresses one of the biggest frustrations with previous LLMs – the black box problem. You'd prompt it, it would give you an answer, and if it was wrong, you had no idea *why* it was wrong or *how* to fix its thought process. Seeing the plan, that's crucial for enterprise adoption. It builds trust.
HostAnd speaking of trust, another major barrier for enterprise adoption has been the notorious "hallucination problem" – the AI just making stuff up. OpenAI claims significant improvements here with GPT-5.4. They're reporting a 33% reduction in factual errors compared to GPT-5.2, and an 18% reduction in overall error rate in responses.
ExpertThat's massive. Hallucinations are the Achilles' heel for anything mission-critical. If you're going to deploy an AI in legal, finance, or consulting, where accuracy is paramount, you simply cannot have it confidently making up facts. A 33% reduction in factual errors is a huge confidence booster. It starts moving AI from "useful but verify everything" to "pretty reliable, but still oversee."
HostSo, we've covered the power, the new capabilities, and the improved reliability. Now, let's go deeper into what that 1-million-token context window *actually* means in practice, beyond just the raw number. You called it a "small library" – that's a great analogy.
ExpertIt really is. Imagine, the LLM's context window is like a human's short-term working memory. If you can only hold three or four ideas in your head at once, you're constantly forgetting and having to re-read. But if you can hold, say, an entire book's worth of information, your ability to connect concepts, understand nuances, and synthesize information skyrockets. That's what this is doing for the AI.
HostAnd that changes the type of tasks it can handle. Previously, you'd have to break down a big report into sections, feed them to the AI, summarize, then feed the summaries to the AI to get a meta-summary. It was this clunky, back-and-forth process. With a million tokens, it can analyze entire code repositories, process hundreds of research papers, or review massive legal contracts, all in one go.
ExpertThis is where we move from automating discrete tasks to automating *entire projects*. Think about a full market research project. You used to have a team of junior analysts spending weeks, if not months, sifting through sales data, market reports, customer surveys. Now, you could theoretically feed all of that raw data into GPT-5.4.
HostAnd it can then ingest and analyze within its context window, identify trends, correlations, anomalies, and then generate a complete PowerPoint presentation or a detailed written report, charts and summaries included. All without the human having to constantly intervene, prompt, and piece things together. No more "back and forth" between human and AI; the AI handles the whole flow.
ExpertThat's a significant streamlining of workflows. It's not just doing one step faster; it's collapsing an entire multi-step process into a single, continuous operation. The efficiency gains there are almost incalculable for knowledge work. But is it too good to be true? Are there any catches?
HostAh, you're hitting on a very important point, and one the research report itself highlights: the "lost in the middle" problem.
ExpertOh, yes. I've heard about this. Even with these massive context windows, models sometimes struggle to recall information if it's buried in the middle of a very long document. They do great at the beginning and the end, but the middle can be a blind spot.
HostExactly. It's like the "serial position effect" in human psychology – we remember the first items on a list (primacy effect) and the last items (recency effect) best, but the stuff in the middle gets a bit fuzzy. Researchers have found that even models designed for long context processing can see performance degrade by over 30% when crucial information is stuck in the middle.
ExpertSo, a million tokens is great, but if the AI misses a critical clause in the middle of a 500-page legal document, that's a problem. OpenAI hasn't explicitly detailed how GPT-5.4 mitigates this, have they?
HostNo, the report doesn't go into detail on that, and it's a critical area of uncertainty for users relying on it for high-stakes analysis. Potential solutions in the field include strategically re-ordering documents to place the most relevant information at the beginning and end, or using more sophisticated retrieval techniques to filter and present only the truly essential context. But whether GPT-5.4 has truly cracked this is still an open question. It means humans still need to be aware of this potential cognitive blind spot.
ExpertThat's a crucial point of skepticism, and it tempers the hype appropriately. Even with all this power, there are still fundamental challenges that need to be addressed. It means human oversight isn't just about ethics; it's about making sure the AI isn't simply "losing" vital information.
HostWhich brings us perfectly to our next big theme: "Agentic AI" is no longer just a technical buzzword. It's now a C-suite problem. This isn't just for the IT department anymore; this is for the CEO, the CFO, the COO.
ExpertAbsolutely. When an AI can execute complex, multi-step business workflows from end to end, that fundamentally changes how you design your business. It's not about deploying an isolated AI tool for a single task; it's about redesigning core business processes around these autonomous agents.
HostWe're talking about creating "digital workers" that can handle entire corporate functions: processing invoices, managing customer service tickets, even doing preliminary financial analysis. This requires a complete rethinking of how work is structured and who – or what – is doing it.
ExpertThe implications for organizational design are profound. If you can automate entire functions, you need fewer people in those functions, or those people's roles shift dramatically. This is no longer just about efficiency; it's about strategic competitive advantage. And it means the major tech players are building the infrastructure for this.
HostTotally. It's not just about OpenAI releasing a powerful model; it's about the platforms that allow enterprises to actually deploy, manage, and govern these agents at scale. Microsoft, with Copilot Studio, is integrating AI agents deeply into its ecosystem – Microsoft 365, Teams, Azure.
ExpertAnd Google Cloud has Vertex AI Agent Builder, Salesforce has Agentforce for sales and customer service, and then you have the established players like ServiceNow and UiPath integrating AI agents into their ITSM and RPA platforms. They're all building the "operating system" for these AI agents.
HostExactly. These platforms are focusing on governance, safety, deep system integrations, and observability. Because when you have thousands of AI agents potentially making decisions and taking actions within your corporate systems, you need to know exactly what they're doing, ensure security, and maintain compliance.
ExpertThat sounds like a whole new level of IT management complexity. But it's essential. Because as powerful as these agents are, they're not fully autonomous, nor should they be. This leads us to the crucial concept of "human-in-the-loop."
HostHuman-in-the-loop, or HITL, is becoming a foundational principle for responsible AI deployment. These agents are designed to execute routine steps independently, but they should pause and request human approval when they encounter uncertainty or a high-impact decision.
ExpertThis is critical. It's about blending the scalability and speed of AI with the nuance, emotional intelligence, and contextual understanding that only humans can provide. It's not about replacement; it's about synergy.
HostAnd accountability. HITL creates a clear line of accountability, ensuring that a human is ultimately responsible for critical actions taken by an AI system. Imagine an AI negotiating a multi-million-dollar contract. You absolutely need a human to sign off on the final terms, even if the AI did 99% of the drafting and analysis.
ExpertIt's about designing workflows where the AI handles the heavy lifting, the data crunching, the pattern recognition, and then presents options or flags issues for human judgment. The most successful organizations won't be aiming for full automation, but for a symbiotic relationship.
HostWhich shifts the focus of work, right? It brings us to the new competitive moat and who wins and loses in this new landscape. Let's look at industries built on repeatable knowledge work and structured processes. Two big ones come to mind: consulting and Business Process Outsourcing, or BPO.
ExpertOh, absolutely. Consulting firms are going to feel this keenly. The traditional model relies on teams of junior consultants spending countless hours gathering and synthesizing information, doing market research, data analysis, drafting reports. AI agents can now perform many of those tasks that used to consume the bulk of a junior consultant's time.
HostSo, consulting teams will likely become smaller and more productive, but what they deliver will change. Clients won't want to pay top dollar for services that can be partially automated. They'll demand deeper industry insight, strategic navigation of complex change – skills that AI can't easily replicate. It elevates the human role to a much higher-level, strategic function.
ExpertAnd BPO, that's another massive one. That industry has historically been built on labor arbitrage – finding cheaper labor overseas to handle routine tasks. AI transforms BPO from a cost-saving play to one driven by technology and intelligence. AI agents can automate data entry, invoice processing, routine customer service.
HostThis leads to the rise of what they're calling "AI-first" BPOs. These aren't just BPOs that *use* AI; they're BPOs whose core value proposition *is* intelligent automation. They handle routine tasks autonomously, freeing up human agents for more complex, empathetic, and high-value interactions. They offer 24/7 service through AI, and they generate data-driven insights.
ExpertIt's a race to adopt these new capabilities, or get left behind. The competitive landscape for service providers is shifting dramatically. It's no longer just about who has the cheapest labor, but who has the smartest AI and the best operational model to deploy it.
HostAnd this fundamentally reshapes corporate knowledge work itself. Roles that are primarily routine, codifiable tasks – data collection, document verification, basic analysis – those are at high risk of automation.
ExpertWhich means a massive shift in skills. Roles that require complex problem-solving, strategic thinking, creativity, and interpersonal skills – those are the ones that will be augmented, not replaced. As AI handles more of the information processing, there's a greater premium on skills like negotiation, leadership, and navigating complex organizational dynamics.
HostIt's a huge shift. The future of knowledge work isn't just humans working; it's humans and their AI "digital colleagues" collaborating. This is going to demand a lot from education and corporate training to re-skill the workforce.
ExpertAbsolutely. It’s not just about job displacement; it’s about job transformation. The nature of human work is evolving, and it’s evolving rapidly.
HostSo, what are the big takeaways from all this, the key insights we should really be mulling over?
ExpertI think first, it's that AI has truly crossed a threshold from being a smart tool to an autonomous agent. The native computer control and the 1-million-token context window are not just improvements; they're paradigm shifts that enable entirely new types of workflows.
HostAnd second, this means AI is now a C-suite concern. It's no longer just an IT project; it's about strategic workflow redesign, rethinking how entire departments operate, and building the necessary enterprise-grade platforms to manage these agents safely and securely.
ExpertThird, the human-in-the-loop isn't a luxury; it's a strategic necessity. For both reliability and accountability, we need to design systems where human judgment remains paramount for high-stakes or ambiguous decisions. It's about a symbiotic relationship, not full replacement.
HostAnd finally, industries built on knowledge work – consulting, BPO – are facing immediate and profound disruption. The competitive moat is shifting from labor cost to AI capability, which means everyone needs to adapt or risk being left behind by "AI-first" competitors.
ExpertIt's a lot to digest. And it raises some really big questions for all of us.
HostFor sure. Like, how quickly will enterprises actually move from just experimenting with AI agents to fundamentally redesigning their core workflows? What are the biggest barriers – technical, cultural, or even regulatory?
ExpertAnd has OpenAI *really* solved that "lost in the middle" problem for those massive context windows, or will that remain a critical vulnerability for complex reasoning tasks? Because that's a huge unknown.
HostAnd for our listeners, what new skills do *you* think will become most valuable in this new world of human-AI collaboration? How should we be preparing ourselves and our teams for this future?