May 30, 2023 · 42m · we-live-to-build
3,000 Client Q&As Fed Into One Bot That Answers Like Him
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Consulting CEO Oleksandr Vysotsky and podcast host Sean Weisbrot examine how artificial intelligence, custom voice cloning, and automated knowledge bots are transforming global business advisory and media workflows. They discuss practical strategies for digitizing executive expertise alongside the technical constraints, safety guardrails, and workforce disruptions created by AI adoption.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Sean holds 42.2% of the talking time here. How this is scored →
Sean as informed peer 4.4 Guest teaching 3.2 Guest disagreement 0.4 Sean pushing back 0.4
speaking balance: gold is Sean, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 31:11 Skepticism on cloning executive decision-makers
Oleksandr firmly rejects the easy assumption that AI clones can replace top executives like Warren Buffett, highlighting the gap between entertainment avatars and real strategic judgment.
Hardest push from Sean ▶ 31:39 Host insists executive clones are fully viableSean pushes back against Oleksandr's skepticism, arguing that if an AI possesses an executive's complete conceptual framework and knowledge base, it can continue running an enterprise.
Biggest teaching moment ▶ 32:30 AI safety boundaries restrict executive decisionsOleksandr educates Sean on how hard-coded safety guardrails cause LLMs to dodge difficult management realities, like firing staff, which inherently impairs high-stakes business decision-making.
Sean holds their own ▶ 36:16 Sean maps coaching framework onto Pi AI architectureSean demonstrates deep tool knowledge by taking Oleksandr's theoretical coaching-versus-consulting distinction and mapping it directly onto the behavioral design of Reid Hoffman's Pi AI.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Sean as informed peer | Guest teaching | Guest disagreement | Sean pushing back | Why |
|---|---|---|---|---|---|---|
| The Vision for AI-Driven Client Support | 3 | 1 | 0 | 0 | Sean introduces Oleksandr and provides context regarding his consulting business and experiments with voice and video cloning. The dynamic is welcoming and collaborative with no conflict. | |
| From Ukrainian Manufacturing to Global Consulting | 2 | 3 | 0 | 0 | Oleksandr details his background founding and selling Ukrainian manufacturing companies right before the 2008 financial crisis. Sean asks open follow-up questions to understand how he anticipated the market downturn. | |
| Exploring Multi-Language AI Voice and Video Tools | 5 | 3 | 1 | 1 | Oleksandr explains the visual awkwardness of early AI video lip-syncing across different languages. Sean demonstrates industry knowledge by comparing this to historical film dubbing in China and introducing Unreal Engine's MetaHuman. | |
| Leveraging CoachVox and 3,000 Client Q&As | 3 | 4 | 0 | 0 | Oleksandr outlines how feeding 3,000 client Q&As into CoachVox can eliminate human consultant drift and overcome timezone delays. Sean listens and validates the operational utility. | |
| Navigating Prompt Engineering and AI Training Challenges | 6 | 3 | 0 | 0 | Sean names specific niche productivity tools such as Pragma and Taskwise.ai while breaking down CoachVox pricing tiers. Oleksandr agrees and explains the necessity of structured prompt training for his staff. | |
| Evaluating Wisdom AI and Celebrity Chatbot Limitations | 4 | 3 | 1 | 0 | Oleksandr shares his disappointing trial of Wisdom AI, specifically testing a Russell Brunson avatar that produced irrelevant answers. Sean acknowledges the evaluation and pivots the discussion to practical production stacks. | |
| Building an End-to-End AI Content Production Pipeline | 3 | 5 | 0 | 0 | Oleksandr provides a masterclass on his end-to-end AI content generation pipeline, combining GPT-4, ElevenLabs voice cloning, and HeyGen avatars. Sean asks clarifying questions about model versions. | |
| Scaling Content Speed and Impact on Creative Jobs | 5 | 2 | 0 | 0 | Sean shares his direct experience using Opus.pro to automate video short creation, which replaced a dedicated human editor. Oleksandr reflects on the broader economic threat to creative workers. | |
| Digitizing Personal Knowledge and AI Self-Cloning | 6 | 1 | 0 | 0 | Sean drives an extended conceptual exploration of digitizing personal knowledge and creating interactive avatars for deceased relatives. Oleksandr expresses excitement about cloning himself to scale his availability. | |
| AI Safety Constraints and Executive Decision-Making Limits | 4 | 6 | 2 | 2 | Sean proposes cloning high-level executives like Warren Buffett, but Oleksandr pushes back by identifying hard limitations in AI decision-making, notably ChatGPT's refusal to recommend firing employees due to internal safety guardrails. | |
| Contrasting Coaching and Consulting with Pi AI | 7 | 4 | 0 | 1 | Oleksandr distinguishes coaching (asking questions) from consulting (giving advice). Sean expertly connects this conceptual split to Inflection AI's Pi bot, explaining its unique conversational and voice interface. |
Statements from this episode (12)
Vysotsky is training an AI coach on 3,000 recorded client Q&As
“For the years I recorded audio and video and articles, and it's about 3000 questions, Q and days from my clients. So I have enough for content to feed to this coach and let's see.”
Vysotsky: AI answers simple consulting questions better than creative human consultants
“Unfortunately sometimes they just misguide our clients because there are some very simple, particular answers to our clients' questions. But if we deal with human beings, we all are creative and we want, want to enhance, we want to add something new. So, and I…”
Vysotsky: Modern software is too complicated without dedicated user training
“Currently it's not enough just to give people a piece of software or some kind of service. You also need to teach them how to use it because every tool we use became Pretty complicated.”
Vysotsky: Only 1% of Google Sheets users likely ever use formulas
“Probably one percent of all the users, they use formulas and a little bit more complicated stuff instead of just putting text into this cells, you know?”
Vysotsky: Wisdom AI's Russell Brunson clone provided stupid and irrelevant answers
“And but unfortunately answers were absolutely stupid. And irrelevant. And this is why I skipped this service.”
Vysotsky: ChatGPT translates foreign languages into English better than human translators
“It translate from foreign languages to English and does it pretty good. Better than most of you know, human translators.”
Vysotsky: ElevenLabs' text-to-speech sounds significantly more realistic than Google or Amazon
“There are some basic services like Google, Amazon, and everything. They all have this text to speech things, but 11 Labs creates a really very, very real sound. It sounds like a human being.”
Vysotsky: Implementing AI video production is 20x faster and 10x cheaper
“20 times faster and 10 times cheaper.”
Weisbrot replaced video editor with Opus, creating 15 clips in 10 minutes
“I used to have a guy, unfortunately, I let him go because Opus makes him unnecessary. I was paying him, it would take him about a Two or three hours to make a short. And now I get 15 in 10 minutes.”
Vysotsky: ChatGPT's safety guardrails prevent it from recommending necessary employee terminations
“When you deal with chat GPT and through him this kind of situations, he tries to avoid firing as much as it can. It's impossible, you know, almost impossible to make him you know, type something like, okay, you have to fire this person. And I believe It happen…”
Vysotsky: Coaching guides clients through questions, whereas consulting provides direct answers
“The first one is coaching and the second one is consulting. So there are two different ways because when it comes to coaching mode, you ask questions. And by changing these questions, you help your client to find the way to kind, to find the right decision. Bu…”