Question 1

Are large language models the path to AGI, or an off-ramp?

Accepted Answer

_LLMs are powerful and useful, but scaling them may be a detour rather than the road to general intelligence._

> I don't think LLMs are AGI, nor that fully scaling up LLMs is going to get you to AGI. I sort of agree with Yann LeCun that on the path to AGI, LLMs are an off-ramp. I still think we can get there by 2029 or even sooner, because the pace of progress is just insane and there are a lot of other techniques now being tried at smaller scale.

Source: [Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052826/)

Question 2

Can AI transcription tools introduce bias based on employee names?

Accepted Answer

_Meeting transcription and AI-driven contribution tracking may disadvantage non-Western names._

> Tools like Granola, Read.ai, Fireflies, and Avoma now routinely track talk time and participation, and firms like Amazon and Meta are folding AI-surfaced contribution into how they review people.

Source: [If Your Name Isn't Western, AI Could Cost You The Promotion.](https://www.forbes.com/sites/lutzfinger/2026/05/25/if-your-name-isnt-western-ai-could-cost-you-the-promotion/)

Question 3

How are AI agents different from RPA (robotic process automation)?

Accepted Answer

_Agents add reasoning where RPA's rule engine stops._

RPA is a rule engine. It automates structured, low-skilled tasks where every decision is essentially zero or one, which is why UiPath and Automation Anywhere built real businesses but hit a ceiling: a script either breaks or runs, and it cannot handle uncertainty or unstructured data. Agents add a language model on top, so they can read an email's style, cope with ambiguity, and self-correct. Cherry's portfolio company Octomind does this for front-end testing with a self-healing loop that notices when a test goes wrong and tries another path, something a rule engine cannot do. The catch: you still have to give an agent clear boundaries, or it goes rogue.

— [The Edge: “AI Agents: The Future of Work?”](https://www.lutzfinger.com/transcript/2024-12-12-ai-agents-the-future-of-work/) · _Podcast_

Question 4

How are AI agents rewriting the internet and changing shopping?

Accepted Answer

_OpenAI and other LLM providers are becoming the dominant interface for online discovery._

> OpenAI's ambition to become the universal interface to what we knew as the Internet isn't surprising. If anything, they might be late. Remember Perplexity's 'Shop Like a Pro'? Their goal was to handle all of shopping — from research to payment.

Source: [AI Agents - The Internet Is Being Rewritten - Survey Users Vs Brands?](https://www.forbes.com/sites/lutzfinger/2025/07/24/ai-agentsthe-internet-is-being-rewrittensurvey-users-vs-brands/)

Question 5

How are jobs changing due to AI?

Accepted Answer

> The word 'engineer' still appears in job postings. So does 'product manager.' But the people being hired for those roles bear little resemblance to the people being let go.

Source: [Meta Laid Off 8,000 And Launched AI - Why Jobs Need Different Skills](https://www.forbes.com/sites/lutzfinger/2026/04/25/meta-laid-off-8000-and-launched-aiwhy-jobs-need-different-skills/)

Question 6

How can a cancer cell be made to produce its own drug?

Accepted Answer

_The aim is a therapy produced by the tumor, against the tumor, in the tumor, so potent drugs stay local._

> We created the first genetic switches made out of DNA that only turn on in cancer cells. You inject them and they go everywhere, but they only flip on inside a cancer cell, like a light switch. We only get into about 1% of the cancer cells, so the cell produces a message, a smoke signal, that brings the immune system to attack the rest, like a domino chain reaction. The hard part is keeping all of it inside the tumor so you avoid off-target effects.

Source: [Tomorrow's Medicine, with Cyriac Roeding (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052126/)

Question 7

How can a large language model design DNA for new cancer drugs?

Accepted Answer

_DNA is a four-letter text, which makes sequence design a next-token problem, but only with the right data and a tight feedback loop._

> What is DNA? It's text. The alphabet of DNA has four letters and they keep permutating, so it's a perfect problem for an LLM, if it's trained on the right data. The general models are only okay-ish for DNA. There's one trained specifically on DNA, Evo 2 by the Arc Institute, the best out there, but still 99% of what comes out is useless. So we built a proprietary Oracle to predict what is not useless, feed it to the wet lab, and the outliers go back into the model. Every spin of the wheel, it gets smarter.

Source: [Tomorrow's Medicine, with Cyriac Roeding (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052126/)

Question 8

How can AI support frontline workers who aren't sitting at computers all day?

Accepted Answer

_Think voice-first interfaces and knowledge retrieval, not just desktop tools._

Frontline workers move around and might not have time to sit at a laptop, but they could have a quick second to speak or type a question. You could support them with an AI interface that answers questions in their language, translates into technical knowledge, and searches knowledge databases. The key is making sure answers are right. Don't just put ChatGPT on your FAQs. You need human in the loop at the beginning, testing a lot to ensure frontline workers are happy with the output before rolling it out widely.

— [You asked, we answered](https://www.lutzfinger.com/transcript/2023-07-28-you-asked-we-answered/) · _The Edge_

Question 9

How can educators use AI clones to scale their teaching?

Accepted Answer

_Creating an AI version of yourself to teach online courses at scale._

> I created the Lutz-Clone, an AI version of myself designed to teach my online certificate about AI with eCornell. This clone looks like me, but more importantly, it contains my knowledge and teaching style.

Source: [AI In Education — Up-level Your Teaching With AI By Cloning Yourself](https://www.forbes.com/sites/lutzfinger/2025/03/31/ai-in-educationup-level-your-teaching-with-ai-by-cloning-yourself/)

Question 10

How can startups use large language models to improve their products?

Accepted Answer

_Simplify interfaces and gain user insights through language_

As a startup founder, think about which interface will be changed by language input. Whenever users needed to use their mouse or learn complex clicking and dragging, you suddenly have the ability to let them do it via language. This simplifies the interface dramatically. Additionally, language prompting gives you more feedback. Before, you'd create an interface and see how users clicked buttons and moved around without really knowing their intentions. Now with language interfaces, you actually see what they're searching for and asking for. This gives you valuable input for building your product.

— [AI as an interface](https://www.lutzfinger.com/transcript/2023-04-24-ai-as-an-interface/) · _The Edge_

Question 11

How do biases in narrow AI systems create real-world harm?

Accepted Answer

_AI systems enforce human prejudices at massive scale._

Narrow AI amplifies existing biases at scale. In parole decisions, computers were meant to remove human bias, but when fed biased data, they kept more minorities in jail. Loan application systems with few women in training data couldn't correctly handle female applicants. The stop and frisk program in New York used a computer to select precincts, but it was racially unjust. When you remove skin color from the data, the system became even more skewed. There's a logical data science explanation for this.

— [Bias in, bias out: Addressing the impact of bias in AI](https://www.lutzfinger.com/transcript/2023-06-07-bias-in-bias-out-addressing-the-impact-of-bias-in-ai/) · _The Edge_

Question 12

How do venture firms use AI in their own operations?

Accepted Answer

_AI in sourcing, diligence, portfolio construction, and founder support._

> AI — it's who controls the data and who builds products that businesses and people actually use.

— [Stargate: The $500 Billion Message To China — The Future Of AI Regulation](https://www.forbes.com/sites/lutzfinger/2025/01/22/stargate-the-500b-message-to-china---the-future-of-ai-regulation/) · _Forbes_

Question 13

How do you measure quality in healthcare data?

Accepted Answer

_Quality metrics for healthcare ML._

> Knowing quality is therefore of utmost importance. Three broad measurement areas in today's healthcare: (1) Metrics about quality of care — related to treatment effectiveness and patient satisfaction. (2) Metrics about need and utilization of care — gauging health requirements of a population and how often one uses what type of service. (3) Metrics about value-based healthcare — they measure disease prevention rather than disease…

— [Healthcare And Data - How To Measure Quality](https://www.forbes.com/sites/lutzfinger/2023/06/27/healthcare-and-datahow-to-measure-quality/) · _Forbes_

Question 14

How do you spot social media bots?

Accepted Answer

_Detection heuristics._

> Figure 6-4. Subset of a social network, created using LinkedIn Labs Since bot networks will first create their own group by linking to one another before they start to infiltrate the rest of the network, especially in the beginning, you will see a highly dense network with very low cohesion to the rest of the network. However, those measures using network topology only work in the beginning as the bot network is formed. Once the bots…

— [Finger Dutta Ask Measure Learn (1)](https://www.lutzfinger.com/book/) · _Book — Ask, Measure, Learn_

Question 15

How does AI change healthcare?

Accepted Answer

_AI applications, value-based care, regulation._

> I just got back from the JP Morgan conference. A few learnings that apply to any industry, not just healthcare:Humans + AI will change the industry. Not AI alone.Cash flow matters. Not AI hype.The real AI revolution is operational.One of the most hyperbolic promises in healthcare in the last month was the claim that: "AI is going to design our drugs and will replace researchers." In my view, this is not true.

— [Healthcare's AI Lesson: Autocomplete Isn't Understanding](https://www.forbes.com/sites/lutzfinger/2026/01/16/healthcares-ai-lesson-autocomplete-isnt-understanding/) · _Forbes_

Question 16

How does AI change the way we interact with computers compared to previous interfaces?

Accepted Answer

_From punch cards to conversation: AI as natural interface_

AI enables me to communicate with machines using natural language instead of learning complex commands or navigating graphical interfaces. Previously, we evolved from punch cards to MS-DOS commands to graphical interfaces with mice and clicking. Each step made things more intuitive, but I still had to search and learn the system's rules. Now with large language models, I can simply describe what I want in my own words. The AI understands my intent and produces output, even if I have no idea how it works technically. This is revolutionary because humans naturally communicate through language.

— [AI as an interface](https://www.lutzfinger.com/transcript/2023-04-24-ai-as-an-interface/) · _The Edge_

Question 17

How does Catalog ensure certain workflows always follow the same steps when using LLMs?

Accepted Answer

_A rules engine ensures critical workflows always execute consistently._

We built a workflow engine underneath our LLMs. The LLM generates a workflow on the fly every time, but users can design deterministic journeys inside our system using workflows (like Zapier but internal). If a request matches a workflow definition, it triggers that workflow every time instead of a dynamically generated journey. It's no code, so users can do it themselves. First, the LLM does a workflow lookup to see if there's something you always want to do, then for the rest it goes into free flow logic.

— [Large Language Models: one interface to rule them all?](https://www.lutzfinger.com/transcript/2023-08-15-large-language-models-one-interface-to-rule-them-all/) · _The Edge_

Question 18

How does Catalog prevent large language models from hallucinating when retrieving enterprise information?

Accepted Answer

_Stopping hallucinations by feeding LLMs the right information first._

The way to avoid hallucination is to stop the model from reaching into the model layers for information. If you give it structured or unstructured information in raw format and then ask it to summarize or list, it gives you exactly what you provided. Without that context, it makes things up because it's doing probabilistic retrieval from its training. This is retrieval augmented generation. You get the information first, augment the prompt with it, then generate the response based on that specific context.

— [Large Language Models: one interface to rule them all?](https://www.lutzfinger.com/transcript/2023-08-15-large-language-models-one-interface-to-rule-them-all/) · _The Edge_

Question 19

How does Catalog use large language models to simplify enterprise software access?

Accepted Answer

_One interface to access all your enterprise tools at once._

Catalog uses LLMs to create a unified interface for all enterprise software. Instead of navigating multiple systems, employees enter a single prompt. The LLM understands the query, identifies which data repository holds the answer (like Workday for payslips), converts the request into an API call, performs deterministic retrieval, and returns the result. We don't use LLMs for data retrieval itself, only for reasoning and synthesis. This compresses workflows from multiple surfaces and ten to fifteen clicks down to one query and two or three clicks.

— [Large Language Models: one interface to rule them all?](https://www.lutzfinger.com/transcript/2023-08-15-large-language-models-one-interface-to-rule-them-all/) · _The Edge_

Question 20

How does LegalOS use LLMs to make legal knowledge more accessible?

Accepted Answer

_LLMs translate legal knowledge into something everyone can use._

LegalOS originally built a no-code visual diagramming tool that lets legal experts encode their knowledge in a structured, machine-readable way. It's like a programming language for law. In recent months, we've pivoted to use LLMs much more heavily to make that legal knowledge accessible. LLMs provide a higher level of abstraction for automations. They let us map open-ended legal requests to structured tasks without requiring someone to manually encode every single rule. The LLM acts as an interface or translator, so humans don't have to understand legalese directly but can still interact with legal text.

— [Will LLMs scale and revolutionize the legal space?](https://www.lutzfinger.com/transcript/2023-06-15-will-llms-scale-and-revolutionize-the-legal-space/) · _The Edge_

Question 21

How has natural language processing changed since 2015, and why does that matter for legal tech?

Accepted Answer

_From counting words to real understanding in less than a decade._

I quit my PhD in 2015 and co-founded a natural language processing company. Back then, the state of the art was basically counting words in a glorified way. You had an idea of what the user might be talking about, but you couldn't map it to something with certainty. You had to collect your own data and train specialized models to do one narrow task. LLMs changed that. You can use foundational models out of the box without fine-tuning. They solve many tasks already. Understanding text and mapping open-world requests to closed sets of tasks became way more accessible.

— [Will LLMs scale and revolutionize the legal space?](https://www.lutzfinger.com/transcript/2023-06-15-will-llms-scale-and-revolutionize-the-legal-space/) · _The Edge_

Question 22

How is AI changing e-commerce?

Accepted Answer

_Fine-tuned models, attribution, category death._

> To explore this question, I looked at one industry: e-commerce. In my online certificate programs, I've explored AI use cases across many industries such as healthcare, media, coffee, finance and e-commerce. While healthcare tends to be slowest to adopt AI, e-commerce consistently leads.

— [ShopTalk 2025: Just 1% Of Vendors Use Generative AI To The Max](https://www.forbes.com/sites/lutzfinger/2025/03/21/state-of-genai--only-1---shoptalk-2025-in-numbers/) · _Forbes_

Question 23

How is AI changing education?

Accepted Answer

_What AI does for and to students, teachers, and curricula._

> AI agents will transform every industry — and one industry I know well is education. So, how will AI agents change how we teach and learn? To explore this question, I created the Lutz-Clone, an AI version of myself designed to teach my online certificate about AI with eCornell.

— [AI In Education — Up-level Your Teaching With AI By Cloning Yourself](https://www.forbes.com/sites/lutzfinger/2025/03/31/ai-in-educationup-level-your-teaching-with-ai-by-cloning-yourself/) · _Forbes_

Question 24

How is DNA being treated as a programmable language?

Accepted Answer

_DNA as text that can be read and written using AI models._

> For fifty years, biology was something we could only read. We are now starting to write it. DNA is next. Models like Evo 2 treat the genome as what it actually is, a language with a four-letter alphabet.

Source: [DNA Is Becoming Programmable. Curing Cancer With AI.](https://www.forbes.com/sites/lutzfinger/2026/05/25/dna-is-becoming-programmable-curing-cancer-with-ai/)

Question 25

How is the Amazon homepage dying?

Accepted Answer

_Retail unbundling, AI assistants as the new aisle._

> The days of big online retail as we know it are over. Amazon, Walmart, Best Buy, and many others will all need to reinvent themselves as e-commerce changes. Once again, the homepage is dead. What will this mean for the big brands?

— [The End Of Amazon, Walmart, Best Buy? AI-Driven Retail Unbundling](https://www.forbes.com/sites/lutzfinger/2025/06/30/the-end-of-amazon-walmart-best-buy---ai-driven-retail-unbundling/) · _Forbes_

Question 26

How is the e-commerce industry changing due to AI?

Accepted Answer

> Every layer is collapsing into the one thing brands actually own, their shopper-behavior data and the fine-tuned AI it makes possible.

Source: [The State Of E-Commerce: Every Layer Of Discovery Changed At Once](https://www.linkedin.com/pulse/state-e-commerce-finetuned-ai-every-brand-lutz-finger-gs5bf/)

Question 27

How is the internet changing due to AI?

Accepted Answer

> The future belongs to AI assistants like ChatGPT, Claude and Perplexity.

Source: [AI Is Changing The Rules Of Web Traffic — A GEO Playbook For Online Retailers](https://www.forbes.com/sites/lutzfinger/2025/08/20/generative-engine-optimization-geo-a-playbook-for-online-retailers/)

Question 28

How many e-commerce vendors are actually using generative AI effectively?

Accepted Answer

_Despite hype, adoption of generative AI in business remains minimal._

> I analyzed 700 vendors at ShopTalk — the biggest e-commerce conference of the year — in regards to generative AI. The results are surprising. Just 1% of vendors use generative AI to the max.

Source: [ShopTalk 2025: Just 1% Of Vendors Use Generative AI To The Max](https://www.forbes.com/sites/lutzfinger/2025/03/21/state-of-genai--only-1---shoptalk-2025-in-numbers/)

Question 29

How might AI impact creative teams?

Accepted Answer

> AI Is Average And That's A Problem The core problem is that AI, by design, is average. It's spectacular at pattern recognition and synthesis

Source: [AI Cuts Creative Teams: What Companies Must Learn About AI Leadership](https://www.forbes.com/sites/lutzfinger/2026/02/13/ai-cuts-creative-teams-what-companies-must-learn-about-ai-leadership/)

Question 30

How should businesses prepare for AGI, and what comes after chatbots and agents?

Accepted Answer

_The practical move is to avoid overcommitting and to build toward agents with persistent memory rather than one-shot chatbots._

> The right answer is not to overcommit to any particular technology at a particular moment, because if the singularity idea is at all real, we're all going to have to pivot and re-pivot. What's coming is agents with persistent memory that reason on that memory and build their own world model as they go. An agent that remembers everything it ever did feels less like a query loop and more like a personal assistant you actually have a relationship with.

Source: [Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052826/)

Question 31

How should executives decide where to apply AI?

Accepted Answer

_Start with the problem, not the tool, and the right use of AI emerges on its own._

> You don't start with the tool, you start with the problem, and then it naturally emerges which tools are best to use. If you want a house, you start with wanting the house, then find the materials and skilled people. You don't start with a screwdriver and ask what you could build with it. AI is like suddenly having a bazooka in your toolbox: it's incredibly powerful, but you still need to aim it and know exactly what to do with it.

Source: [Tomorrow's Medicine, with Cyriac Roeding (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052126/)

Question 32

How will AI change jobs?

Accepted Answer

_Disruption, enablement, and the skills that matter._

> Microsoft offered voluntary retirement to thousands of long-tenured staff. Twenty thousand jobs gone in seven days, from two of the most profitable technology companies in history. The news call it an AI replaces humans. This storyline is wrong.

— [Meta Laid Off 8,000 And Launched AI - Why Jobs Need Different Skills](https://www.forbes.com/sites/lutzfinger/2026/04/25/meta-laid-off-8000-and-launched-aiwhy-jobs-need-different-skills/) · _Forbes_

Question 33

If agents are the future, why won't one giant model just do everything?

Accepted Answer

_The moat is the workflow, not a bigger model._

Because the durable approach is the opposite: smaller, faster, specialized models orchestrated together, not one ever-bigger general model. Even OpenAI orchestrates the long tail of consumer search with smaller models, because a big model would be far too slow. For founders this is the key warning: if your agent's next step is something generic, like deciding whether a calculator is needed, that is the long tail, and the big models will absorb it, just as Google eventually swallowed the specialized search engines. The defensible business is one where the founder knows the specific stepwise workflow of a domain and chains the right models and rules to serve it.

— [The Edge: “AI Agents: The Future of Work?”](https://www.lutzfinger.com/transcript/2024-12-12-ai-agents-the-future-of-work/) · _Podcast_

Question 34

Is AGI near?

Accepted Answer

_Lutz's stance on the AGI timeline._

> Both have been hot topics in AI, sparking both fear and excitement. Sam Altman predicted AGI would arrive in 2025, and Elon Musk 2026. Both are more about hype than reality. In 2025, there will be no AGI but large language models will find their "killer app." What are AGI and the Singularity?

— [No AGI But A 'Killer App' — 2025 AI Prediction](https://www.forbes.com/sites/lutzfinger/2025/01/03/no-agi-but-a-killer-app2025-ai-prediction-110/) · _Forbes_

Question 35

Is AI hurting people with non-Western names?

Accepted Answer

_Concrete bias outcomes in hiring and promotion._

> The junior analyst who keeps pushing the conversation forward finally gets seen, even though she never had the corner office. Tools like Granola, Read.ai, Fireflies, and Avoma now routinely track talk time and participation, and firms like Amazon and Meta are folding AI-surfaced contribution into how they review people.

— [If Your Name Isn't Western, AI Could Cost You The Promotion.](https://www.forbes.com/sites/lutzfinger/2026/05/25/if-your-name-isnt-western-ai-could-cost-you-the-promotion/) · _Forbes_

Question 36

Is Google search dying?

Accepted Answer

_What changes when LLMs become the search interface._

> Back in the 1990s, people said "All the information is out there at the internet... we just need to find it." That was the time of search engines — AltaVista, Yahoo! Search, Google, Lycos, WebCrawler, Excite. Even Disney tried its hand with Go (1998). New Technology — New Search: Large Language Models Build AI Search.

— [2025: The Year Of Search But Not The End Of Google](https://www.forbes.com/sites/lutzfinger/2025/01/03/2025-the-year-of-search-but-not-the-end-of-google/) · _Forbes_

Question 37

Is OpenAI going bankrupt?

Accepted Answer

_Lutz's take on OpenAI economics and moats._

> OpenAI — "Experts predict bankruptcy" — was attention-grabbing, a deeper look reveals that the company has simply seen a drop in its users. Should you fear ChatGPT will disappear? No, quite the opposite.

— [OpenAI Isn't Going Bankrupt, But It Has A Business Model Problem](https://www.forbes.com/sites/lutzfinger/2023/08/18/is-openai-going-bankrupt-no-but-ai-models-dont-create-moats/) · _Forbes_

Question 38

Is SaaS dead due to AI?

Accepted Answer

> The interface is. SaaS is not dead. The interface is.

Source: [SaaSpocalypse Is Dead - The Future Of SaaS Is SaaS](https://www.forbes.com/sites/lutzfinger/2026/04/24/saaspocalypse-is-deadthe-future-of-saas-is-saas/)

Question 39

Should AI be regulated?

Accepted Answer

_Lutz's view on AI regulation._

> But regulation doesn't happen in a vacuum. The EU has regulated that gender should not be used in AI models — even though this may lead to less effective risk distribution and higher overall costs. As a society, we may be willing to pay that price, but what if other societies are not? Data is the lifeblood of AI.

— [Emmanuel Macron On AI: We Are Regulating Ourselves Out Of The Market](https://www.forbes.com/sites/lutzfinger/2024/10/22/emmanuel-macron-is-right-about-ai-over-regulatingi-have-no-doubt/) · _Forbes_

Question 40

Should you trust OpenAI blindly?

Accepted Answer

_Vendor risk and architectural diversification._

> Stargate's $500B investment — message to the competitors — meant to intimidate potential challengers with overwhelming financial backing. AI is extremely dangerous; only a select few should be allowed to develop it — OpenAI's storyline to use safety concerns as justification for a regulatory setup that creates barriers to entry and gives them more monopolistic power. From now on, models should respect IP rights — OpenAI ignored IP…

— [Is OpenAI A Victim — DeepSeek And IP Rights — A Playbook For Leaders](https://www.forbes.com/sites/lutzfinger/2025/01/29/is-openai-a-victim---deepseek-and-ip-rights---a-playbook-for-leaders/) · _Forbes_

Question 41

What about deepfakes?

Accepted Answer

_How society will adapt to synthetic media._

> As AI lowers the barriers to content creation, abuse is also becoming more feasible. Deepfakes — AI-generated images, videos and audio designed to mimic real people — pose serious risks. Already in 2022, I demonstrated how easily they can be created. This increases the risk of misinformation, identity theft and political manipulation.

— [Goku, A Chinese Open-Source Model, Challenges OpenAI And The U.S.](https://www.forbes.com/sites/lutzfinger/2025/02/11/goku---a-chinese-open-source-model-challenges-openai-and-the-us/) · _Forbes_

Question 42

What are the core limitations of transformers and backpropagation for building AGI?

Accepted Answer

_The train-then-freeze design of today's models is the wrong shape for a mind that keeps learning._

> Because of the underlying backpropagation algorithm, learning and inference are separate phases. You train a model, you freeze it, and then you use it. If you try to fill the network up with knowledge and keep going, at some point you get catastrophic forgetting: adding a little new knowledge causes it to forget a lot of old knowledge. These systems know a lot, but they don't reorganize themselves in real time the way parts of the human brain do.

Source: [Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052826/)

Question 43

What are the four approaches to addressing bias in AI models?

Accepted Answer

_Four practical methods to reduce AI bias in production systems._

First, identify the bias by testing features that shouldn't have impact, like running loan applications 100 times to see gender differences. Second, help users understand potential bias through UX, like asking specifications when someone requests an image of a nurse. Third, correct the data through prompting, prompt tuning, or fine-tuning with better examples. Fourth, create guardrails or constitutions that let AI critique itself based on rules. Each approach has limitations, and changing underlying biased datasets in large models remains hard.

— [Bias in, bias out: Addressing the impact of bias in AI](https://www.lutzfinger.com/transcript/2023-06-07-bias-in-bias-out-addressing-the-impact-of-bias-in-ai/) · _The Edge_

Question 44

What could be the deciding factor in the AI race?

Accepted Answer

> But the real value may sit in the one thing OpenAI, Anthropic, and Gemini have not yet made central: your own eval data.

Source: [The Missing Moat In AI: Your Eval Data](https://www.forbes.com/sites/lutzfinger/2026/05/26/the-missing-moat-in-ai-your-eval-data/)

Question 45

What did Jim Allison's checkpoint inhibitors miss, and how does early aim to complete the picture?

Accepted Answer

_Releasing the immune system's brake was a Nobel-winning breakthrough, but a brake release alone does not move the car._

> Checkpoint inhibitors are the biggest cancer innovation in the last 20 years, and Jim Allison won the Nobel Prize in 2018 for the insight that you first have to take the foot off the brake of the immune system. But even the best of them have on average a 35% response rate. The problem is, if you take your foot off the brake but never push the gas, the car doesn't move. You also have to stimulate the immune system, and we do that locally at the tumor. That is the second missing piece of the puzzle.

Source: [Tomorrow's Medicine, with Cyriac Roeding (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052126/)

Question 46

What does 'AI as an operating model' mean?

Accepted Answer

_Why the winners redesign work around AI rather than just adopting tools._

> AI — it's who controls the data and who builds products that businesses and people actually use.

— [Stargate: The $500 Billion Message To China — The Future Of AI Regulation](https://www.forbes.com/sites/lutzfinger/2025/01/22/stargate-the-500b-message-to-china---the-future-of-ai-regulation/) · _Forbes_

Question 47

What does the eCornell AI certificate cover?

Accepted Answer

_The five-course program structure._

> While AI can enhance scalability, human involvement remains essential. What do students in my eCornell certificate tell me after the course? AI is great at summarization and coding. An AI can help being a dynamic checklist and memory.

— [Overhyped: Manus From Monica — AI Made In China](https://www.forbes.com/sites/lutzfinger/2025/03/10/overhyped-manus-from-monicaai-made-in-china/) · _Forbes_

Question 48

What is 2025 for search?

Accepted Answer

_Predictions about search changing in 2025._

> It might sound like a throwback to the 1990s, but technology has transformed the way we search for and interact with information. If past patterns hold true, 2025 will be the year of search. Will it be the end of Google?

— [2025: The Year Of Search But Not The End Of Google](https://www.forbes.com/sites/lutzfinger/2025/01/03/2025-the-year-of-search-but-not-the-end-of-google/) · _Forbes_

Question 49

What is a data product?

Accepted Answer

_Definition and how to build one._

> Now, to be a good product developer, you need to think about self-referential loops within your products. How to get signal, how to get feedback from your users so your AI, or your data science, can leverage that information to make the product work better. So that moved a little bit within it. The innovation, and the inspiration about, and the curiosity about how to use data in novel manners, that's still data science, because that's…

— [Unit 3 Full Transcript](https://www.lutzfinger.com/ecornell/) · _eCornell course_

Question 50

What is an AI agent, really?

Accepted Answer

_Not a sci-fi character, a workflow that chains steps._

An agent is not a sci-fi character and it is not one giant model. It is a piece of code that takes a defined task, breaks it into steps, and lets you act on software through language. The core idea is chaining: instead of asking a model to "improve revenue," you break the goal into step one, step two, step three, exactly the way a manager hands work to a junior employee. Tools like LangChain popularized this. So an agent is less a breakthrough model than a workflow that combines models, tools, and sometimes plain deterministic code, each doing one step.

— [The Edge: “AI Agents: The Future of Work?”](https://www.lutzfinger.com/transcript/2024-12-12-ai-agents-the-future-of-work/) · _Podcast_

Question 51

What is BioAI?

Accepted Answer

_AI applied to biology — new molecules, new medicines._

> AI, Roeding said, is like suddenly having a bazooka in your toolbox. That means though, you still have to aim, and you still have to pull the trigger.

— [DNA Is Becoming Programmable. Curing Cancer With AI.](https://www.forbes.com/sites/lutzfinger/2026/05/25/dna-is-becoming-programmable-curing-cancer-with-ai/) · _Forbes_

Question 52

What is Generative Engine Optimization (GEO)?

Accepted Answer

_Why being discoverable to AI assistants is the new SEO._

> For brands a new playbook is emerging: AEO. AEO is just the beginning.

— [Study Shows LLM Conversion Rate Is 9x Better — AEO Is Coming](https://www.forbes.com/sites/lutzfinger/2025/06/19/study-shows-llm-conversion-rate-is-9x-better---aeo-is-coming/) · _Forbes_

Question 53

What is predictive coding and why does it matter for AGI?

Accepted Answer

_Changing the learning rule, not just the architecture, may be what unlocks continual learning._

> There's something called predictive coding which more closely resembles how the brain works. On smaller neural nets, predictive coding networks are better at continual learning: they don't have as much catastrophic forgetting, each neuron can learn asynchronously and independently, and you can do inference and learning in the same phase. The need for hyperscaler server farms is less, because the learning is asynchronous.

Source: [Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052826/)

Question 54

What is the 'Ask, Measure, Learn' framework?

Accepted Answer

_The framework from Lutz's book._

> "Ask Measure Learn" is a non-technical guide shows you how to extract significant business value from big data. It is a system that helps you ask the right questions, measure the right data, and then learn from the results. The framework was originally devised to help governments and NGOs sift through volumes of data.

— [Ask Measure Learn](https://www.lutzfinger.com/book/) · _Book — Ask, Measure, Learn_

Question 55

What is the 'billion dollar barrier' in delivering gene therapies?

Accepted Answer

_Getting genetic cargo to a tumor, and not to the liver, has blocked the field for 15 years._

> Somehow you have to get this DNA to the tumor, and that's where lipid nanoparticles come in. You've probably already had them in a COVID vaccine. The problem is that within 10 minutes they're in your liver and out of your body, because the liver flushes out everything that doesn't belong. So to treat any cancer other than liver cancer you need a completely different lipid nanoparticle. For 15 years that has been called the billion dollar barrier.

Source: [Tomorrow's Medicine, with Cyriac Roeding (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052126/)

Question 56

What is the book Ask, Measure, Learn about?

Accepted Answer

_Topics covered in the O'Reilly book._

> **Publisher:** O'Reilly Media **Published:** Feb 1, 2014

— [Ask Measure Learn](https://www.lutzfinger.com/book/) · _Book — Ask, Measure, Learn_

Question 57

What is the competitive advantage of LLMs for business?

Accepted Answer

_How enterprises can build durable AI advantage._

> AI-driven businesses have a combination of either a model moat, a data moat, or the brand moat. The new LLMs, like OpenAI's model, can give an advantage in all areas, but it is by no means a given. Since OpenAI launched ChatGPT, large language models have entered the fantasies of founders worldwide. Every day about a dozen new startups announce their AI-enabled feature.

— [What Is The Competitive Advantage Of LLMs Like ChatGPT For Your Business? Three Takeaways.](https://www.forbes.com/sites/lutzfinger/2023/04/04/what-is-the-competitive-advantage-of-llms-like-chatgpt-for-your-business-three-takeaways/) · _Forbes_

Question 58

What is the core problem with today's cancer drugs?

Accepted Answer

_Even the best cancer drugs are imprecise, and the imprecision forces a dose that limits how well they work._

> The number one problem with cancer drugs is systemic toxicity. When you dose someone, the drug doesn't only work on the tumor, it also works on healthy cells. That's why your hair falls out in chemotherapy. The top 10 cancer drugs produce 87 billion dollars a year, and on average they give you 12 months more to live and 15 percentage points more survival, which gets you to 49% in total, meaning every second person still dies.

Source: [Tomorrow's Medicine, with Cyriac Roeding (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052126/)

Question 59

What is the difference between AGI, the singularity, and superintelligence?

Accepted Answer

_Three different ideas that boardrooms constantly conflate, with the founder who coined "AGI" drawing the lines._

> A truly general intelligence should be able to figure out how to achieve any goal in any environment. To be an AGI you should be able to pivot beyond whatever you were programmed for: given an unfamiliar environment and an unfamiliar goal, figure out on the fly, with creative imagination and trial and error, how to achieve it. The singularity is a point at which the rate of progress in AI, biotech, nanotech and space becomes so fast it would feel infinite to a legacy human being like us.

Source: [Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052826/)

Question 60

What is the difference between virality and contagiousness?

Accepted Answer

_Stories gain weight as they travel, not uniform infection rates._

Virality is something which is equally infectious at each stage. If I have a story which one out of 10 people like, that story doesn't change. One out of 10 people like it in the future as well. Contagiousness is different. As long as it travels, it gains weight. The more people believe in it, the more trustworthy the story becomes. Even a wrong information becomes more believed the longer it travels. At some point everyone believes it's truth no matter what because so many people believed in it. Forget about virality. The right word is contagiousness.

— [Measuring Social Media What Marketers Need To Know](https://www.lutzfinger.com/transcript/2014-08-22-measuring-social-media-what-marketers-need-to-know/) · _Social Media Examiner_

Question 61

What is the economic risk during the transition to AGI?

Accepted Answer

_The dangerous window may not be an evil superintelligence, but the gap between AGI that can automate jobs and superintelligence that can deliver abundance._

> What worries me is the period after you have the first human-level AGI but before you have a superintelligence that can bring massive abundance. You've got an AGI that can in principle automate almost any job and is crunching through one industry after another, but you don't yet have the superintelligence that can airdrop a molecular nano-assembler in everyone's backyard. What happens to the world economy in that interval? Who gives universal basic income in the Central African Republic?

Source: [Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)](https://ecornell.cornell.edu/keynotes/view/K052826/)

Question 62

What is the hallucination problem with large language models and why does it matter for businesses?

Accepted Answer

_LLMs invent answers when they don't know, risking business credibility_

Large language models will make things up when they don't know the answer. I demonstrated this by asking ChatGPT to summarize a completely fake Cornell website that doesn't exist. It generated a summary saying my course was the most amazing and made very good logical arguments, but it was completely fake. For businesses, this is a huge problem. If a customer asks a shop for something it doesn't have, the model might just make it up and promise red shoes with blinking laces. That's why proprietary data matters. You need to ground the model on your local data.

— [AI as an interface](https://www.lutzfinger.com/transcript/2023-04-24-ai-as-an-interface/) · _The Edge_

Question 63

What is the killer app of AI?

Accepted Answer

_Where the durable AI value will accumulate._

> Both are more about hype than reality. In 2025, there will be no AGI but large language models will find their "killer app." What are AGI and the Singularity? Artificial General Intelligence: an advanced AI that can think, learn, and solve problems across a wide range of tasks, just like a human. The Singularity: the idea of AI surpassing human intelligence, improving itself endlessly, and causing massive, unpredictable changes in…

— [No AGI But A 'Killer App' — 2025 AI Prediction](https://www.forbes.com/sites/lutzfinger/2025/01/03/no-agi-but-a-killer-app2025-ai-prediction-110/) · _Forbes_

Question 64

What is the missing moat in AI?

Accepted Answer

_Eval data as the durable moat._

> What wins the AI race? UX?

— [The Missing Moat In AI: Your Eval Data](https://www.forbes.com/sites/lutzfinger/2026/05/26/the-missing-moat-in-ai-your-eval-data/) · _Forbes_

Q&A

Are large language models the path to AGI, or an off-ramp?

Can AI transcription tools introduce bias based on employee names?

How are AI agents different from RPA (robotic process automation)?

How are AI agents rewriting the internet and changing shopping?

How are jobs changing due to AI?

How can a cancer cell be made to produce its own drug?

How can a large language model design DNA for new cancer drugs?

How can AI support frontline workers who aren't sitting at computers all day?

How can educators use AI clones to scale their teaching?

How can startups use large language models to improve their products?

How do biases in narrow AI systems create real-world harm?

How do venture firms use AI in their own operations?

How do you measure quality in healthcare data?

How do you spot social media bots?

How does AI change healthcare?

How does AI change the way we interact with computers compared to previous interfaces?

How does Catalog ensure certain workflows always follow the same steps when using LLMs?

How does Catalog prevent large language models from hallucinating when retrieving enterprise information?

How does Catalog use large language models to simplify enterprise software access?

How does LegalOS use LLMs to make legal knowledge more accessible?

How has natural language processing changed since 2015, and why does that matter for legal tech?

How is AI changing e-commerce?

How is AI changing education?

How is DNA being treated as a programmable language?

How is the Amazon homepage dying?

How is the e-commerce industry changing due to AI?

How is the internet changing due to AI?

How many e-commerce vendors are actually using generative AI effectively?

How might AI impact creative teams?

How should businesses prepare for AGI, and what comes after chatbots and agents?

How should executives decide where to apply AI?

How will AI change jobs?

If agents are the future, why won't one giant model just do everything?

Is AGI near?

Is AI hurting people with non-Western names?

Is Google search dying?

Is OpenAI going bankrupt?

Is SaaS dead due to AI?

Should AI be regulated?

Should you trust OpenAI blindly?

What about deepfakes?

What are the core limitations of transformers and backpropagation for building AGI?

What are the four approaches to addressing bias in AI models?

What could be the deciding factor in the AI race?

What did Jim Allison's checkpoint inhibitors miss, and how does early aim to complete the picture?

What does 'AI as an operating model' mean?

What does the eCornell AI certificate cover?

What is 2025 for search?

What is a data product?

What is an AI agent, really?

What is BioAI?

What is Generative Engine Optimization (GEO)?

What is predictive coding and why does it matter for AGI?

What is the 'Ask, Measure, Learn' framework?

What is the 'billion dollar barrier' in delivering gene therapies?

What is the book Ask, Measure, Learn about?

What is the competitive advantage of LLMs for business?

What is the core problem with today's cancer drugs?

What is the difference between AGI, the singularity, and superintelligence?

What is the difference between virality and contagiousness?

What is the economic risk during the transition to AGI?

What is the hallucination problem with large language models and why does it matter for businesses?

What is the killer app of AI?

What is the missing moat in AI?

What is the neuro-symbolic approach to AGI behind OpenCog Hyperon?

What is value-based care and how does AI enable it?

What makes autonomous AI agents different from traditional automation tools like RPA?

What mistakes do brands make in AI-driven e-commerce?

What patterns make a workflow 'agentic' rather than just a wrapper around an LLM?

What practical business use cases exist for language model chaining beyond autonomous agents?

What problem did transformers solve for Ultimate AI's multi-language support?

What problems with RAGs still need to be solved before widespread adoption?

What replaces SEO in the LLM era?

What was Ultimate AI's key insight that made customer onboarding faster than competitors?

Where do neuro-symbolic AI systems beat large language models today?

Where does Ultimate AI see its competitive defensibility beyond just AI models?

Who will win in the AI era—those using AI or those redesigning work around it?

Why are AI workflows the real moat?

Why are coding applications seeing the biggest success with AI agents right now?