Mustafa Suleyman on Defining Intelligence Seth Rosenberg 2024-06-18 | Meet Mustafa Suleyman, the man leading Microsoft’s AI efforts.In 2010, Mustafa co-founded and led the application of AI at DeepMind.Later, he became the CEO and founder of Inflection AI, and a venture partner at Greylock in 2022.Mustafa also serves on the board of directors of The Economist and is a senior fellow at the Belfer Center for Science and International Affairs at the Harvard Kennedy School.In this episode, I’m thrilled to discuss with him the transformer architecture of LLMs, how incumbents are scaling compute to meet extensive demand, what Microsoft is doing to make the Copilot UI better, and the current state of GPT models.ABOUT ME: Growing up in Winnipeg, Canada, my summers as a kid were spent trying to emulate entrepreneurs around me to start small businesses. I joined Facebook as a product manager in August 2013, leading the transition of launching Messenger as a standalone app and developing the AI bot platform.At Greylock Partners, I invest in early-stage founders, with a focus on AI, SaaS, and Fintech from Seed-Series B. I believe that AI will create massive opportunities for product-focused founders to build incredible companies with strong moats, which is what this show is all about. Some of my portfolio companies at Greylock Partners include 0x, Tome, Greenlite, Espresso, and Roblox. If you’re a product-driven founder trying to build disruptive companies in this era of AI, here are three opportunities I wrote about: greylock.com/greymatter/seth-rosenberg-product-led-aiBuilding in this space, and you share a similar vision for AI? Please get in touch with me. Connect with us on:1. Mustafa Suleyman- linkedin.com/in/mustafa-suleyman2. Seth Rosenberg- linkedin.com/in/sethgrosenberg00:00:00 Trailer00:00:56 Introduction00:01:31 How Mustafa Sulayman saw AI's future00:04:58 AGI is the AI that we don’t yet have00:08:38 Open Source AI is about to change everything00:13:42 How can startups collect high-quality AI data?00:18:55 What UIs will AI-first companies build?00:20:16 Should we pursue fully autonomous AI?00:25:24 AI's path to 99% accuracy00:27:30 Designing AI for ambiguous problem domains00:32:08 AI Copilot is the ultimate productivity assistant
AI is Going To Be Your Next Coworker | Jared Spataro Seth Rosenberg 2024-08-05 | AI might become your coworker or even your manager!Moving forward, AI will be able to do everything you can now, but with one minor difference—it will likely do it better than you.My guest on today’s episode is Jared Spataro, Corporate VP of AI at Microsoft.He predicts that in five years, copilots will become a skill at least as important as working with people if not more.For incumbents, this marks a shift towards AI-driven re-engineering on the back of near-real-time, and autonomous agents.Jared envisions a “great rewiring” of business processes, driven by AI capabilities in reasoning, planning, and memory.In finance, for instance, traditional practices are giving way to AI-led methods.As barriers to entry dissolve, this is the right time to strike while the iron is hot.
Startups Need To Focus on Data Acquisition over Parameters | Mustafa Suleyman Seth Rosenberg 2024-08-04 | Startups, take note: Constantly chasing larger models is futile. Instead, the moat lies in data acquisition. As an AI founder, you need to look for existing labeled datasets or, even better, design a UI that inherently collects data from user interactions. This will help you create defensibility with high-quality, proprietary data and eliminate the need to depend on large-scale model providers. Imagine smaller, potent models like Microsoft's Phi-3, hitting near GPT-4 levels in a compact, open-source package. With reinforcement learning and model distillation, AI feedback drives fine-tuning, boosting performance without bloating parameters. It's not just about size; it's how you wield your data. In this episode, Mustafa Suleyman of Microsoft emphasizes why founders should prioritize data acquisition and filtering over parameter count.
Reid Hoffman on Why Your Current AI Know-How Will Become Obsolete Seth Rosenberg 2024-07-31 | #ai #tehnology #innovation #podcast Reid Hoffman advises founders to get rid of any preconceived notions about AI and how it should be.You see, the most impressive models today will likely become more sophisticated over time.As compute scales, so will datasets, leading to more fine-tuned models with functionalities beyond our imagination.With GPT-5 underway, Reid is reminded of how each new stack feels much better than its predecessor, thanks to the natural order of compounding loops at scale.You can’t control what the market will throw at you next, but if you want to disrupt the ecosystem, you need to consistently iterate and refine models to meet user needs.In this episode of Product-Led AI, Reid also discusses why recognizing ideas that most people dismiss as foolish is important for founders.
Crypto Shaping Our Politics & Governance | Balaji Srinivasan Seth Rosenberg 2024-07-29 | SEC crackdowns, ETFs, and BlackRock's entry— crypto has seen it all this year.Looking ahead, crypto's influence on US politics will deepen, with lawmakers inevitably forming crypto tribes. And this is a global effect, that world leaders are leveraging to shape their politics.President Bukele of El Salvador has made it clear in his ‘HODL’ tweets that Bitcoin is much more than a store of value.Moving forward, Bitcoin will act as a bottleneck on the money printers of sovereign states, potentially curbing federal rate hikes, and contributing to more stable CPI readings. Balaji Srinivasan, ex-CTO of Coinbase, argues crypto's pivotal role in AI governance will curb excessive money printing.
How LinkedIn, Airbnb, and Meta Keep Winning Seth Rosenberg 2024-07-15 | Building successful networks is incredibly complex.LinkedIn, Airbnb, and Meta have done it with ease.How?For one, differentiation is crucial.People often assume all networks are established and set in stone, as they did before Snapchat took off.Some startups succeed by layering minimal enhancements on say, ChatGPT, and think they’ve built a valuable product.The question, however, remains if the product has durability and defensibility.This involves a deeper understanding of the compounding loops of growth within the network.#linkedin #ai #airbnb #meta #innovation #technology #podcast
The Consequences of Opposing Open Source AI Seth Rosenberg 2024-07-13 | Arguing against open source in the name of safety is shortsighted.Those advocating for halting AI development are doing so purely to line their pockets, says Balaji Srinivas.What they fail to realize is that monopolizing research will stifle progress.In this episode of PLAI, Balaji stresses that his research shows how incumbents can collaborate to fund open-source models together.This would cut the high costs associated with API calls from closed-source models.Collaboration, not monopolization, drives true innovation.
AI Will Create More Jobs On the Back of Crypto Seth Rosenberg 2024-07-12 | Digital blue-collar work will single-handedly level labor arbitrage.Data labelers, who collect training data to refine AI models, earn just over $13/hour.Is it too little?For those in first-world countries, perhaps.But in a country like India, where GDP per capita is less than $3,000/year, $13/hour is a decent income.More Data → More Jobs = Better Training PointsIt’s a win-win.In this episode of PLAI, my guest, Balaji Srinivasan, explains why crypto, not fiat, will become the global payment method for everyone participating in this shift.
Crypto is Leading AI’s Fight Against Web 2 0 Seth Rosenberg 2024-07-12 | AI has posed challenges for the open web, but crypto is leading a revival.Web 2.0 is tightening access to their data against indexing, making it tougher for AI models to succeed.Even the likes of Stack Overflow and Reddit are grappling with AI bots manipulating content and gaming upvotes.Cryptography offers a multitude of solutions for this, for example:- NFT gates ensure access;- Signed webs verify human interaction;- Open blockchains enable comprehensive indexing.The hope is that this will counterbalance AI's impact, ensuring integrity in our interactions.
Crypto Will Take Centerstage in the Governance Layer of AI Seth Rosenberg 2024-07-10 | #crypto #ai #cryptocurrency #investing #politics #podcast In Q1 of this year, Spot Bitcoin ETF inflows stood at $12B.From the SEC’s crackdown to Blackrock’s participation– crypto has seen it all in the past year.In the future ahead, we'll see crypto shaping politics even more.Lawmakers will invariably become a crypto tribe.President Bukele of El Salvador has made it clear in his ‘HODL’ tweets that Bitcoin is much more than a store of value.Balaji Srinivasan, former CTO of Coinbase, believes that as part of its broader role in the governance layer of AI, crypto will keep the money printers at bay.
Balaji Srinivasan on Crypto as an AI Accelerant Seth Rosenberg 2024-07-09 | AI is great at making things up (and very quickly, in large quantities). Crypto excels at verifying what's real. Can crypto's big comeback be as a complementary technology to AI?In this episode, my guest is Balaji Srinivasan, a prolific angel investor, the former CTO of Coinbase, and a former GP at Andreessen Horowitz. He shares his viewpoint on crypto as an accelerant to AI development and adoption. He posits that the dual technologies can be leveraged across finance, employment services. “AI for creation, crypto for monetization, social for distribution," says Balaji. Following the disaster that ensued in 2022— rug-pulls, NFTs crashing, and even Elon’s meme coin fascination, Balaji believes crypto can make a comeback.He discusses Bitcoin ETFs, the risk of opposing Open Source in AI, crypto as money in the age of AI, the resurgence of NFTs, and how cryptography will affect AI job prospects.You can read the transcript here: productledaipod.com/podcasts/is-crypto-an-ai-accelerantABOUT ME: Growing up in Winnipeg, Canada, my summers as a kid were spent trying to emulate entrepreneurs around me to start small businesses. I joined Facebook as a product manager in August 2013, leading the transition of launching Messenger as a standalone app and developing the AI bot platform.At Greylock Partners, I invest in early-stage founders, with a focus on AI, SaaS, and Fintech from Seed-Series B. I believe that AI will create massive opportunities for product-focused founders to build incredible companies with strong moats, which is what this show is all about. Some of my portfolio companies at Greylock Partners include 0x, Tome, Greenlite, Espresso, and Roblox. If you’re a product-driven founder trying to build a disruptive company in this era of AI, please get in touch with me. Connect with us here:1. Balaji S.- linkedin.com/in/balajissrinivasan2. Seth Rosenberg- linkedin.com/in/sethgrosenberg00:00:00 Trailer00:00:58 Introduction00:02:16 How AI and Crypto complement each other00:05:33 AI struggles with dynamic environments00:13:06 How Blockchain is transforming financial markets00:19:18 Why banks won't embrace crypto00:21:54 How crypto can safeguard against AI phishing attacks00:30:11 Crypto simplifying payments using AI00:38:18 How crypto is saving the open web from AI00:40:12 Why crypto is key to the next wave of AI job creation00:45:09 Crypto's role in future governance00:51:41 Three types of Bitcoin states explained00:53:59 How decentralized media is changing the truth01:01:34 The risk of opposing open source in AI
Microsoft’s Balancing Act for More Value Creation | Jared Spataro Seth Rosenberg 2024-06-28 | If you're a $3 trillion company, choosing your battles is a Herculean task.Microsoft’s strategy?Build practical applications after extensive trial and error.In this discussion with Jared Spataro, Corporate VP of AI at Work at Microsoft, he speaks to how the company pushed Copilot to the 330M MAUs using Microsoft Teams.Jared and his team are confident that users who fully grasp Copilot's functionalities can save over 10 hours/month.The Big Tech giant is also razor-focused on customer support.At the core of this effort is the foundational GPT model, designed to optimize throughput and efficiency.You can read a transcript of our conversation and find more about the series at productledaipod.com
Perplexity is Beating Google at Its Own Game | Aravind Srinivas Seth Rosenberg 2024-06-27 | Aravind Srinivas is taking on Google's core business head-on: search.Perplexity is streamlining results for quick, precise answers by merging AI and search indexes. Unlike Google, it isn't tied to ad revenue from clicks.No more endless links, and no more hallucinatory responses. Aravind says, “We don't blame the user for not being able to come up with a good question. Instead, we try to have the AI work for the user to get them to a good point.”And that’s what separates Perplexity from the rest.In this episode of Product Led AI, I’m joined by Aravind as he opens up about chatbots, why he’s always wanted to disrupt search, competing directly with Google, and more. You can read a transcript of our conversation and find more about the series at productledaipod.com
Preserving Relationships & Memories with Conversational AI - Replika Seth Rosenberg 2024-06-26 | Launched in 2017, Replika.ai origin story stems from a mission to develop advanced conversational AI.Initially conceived as a memorial bot to sustain conversations with a departed friend, the concept was groundbreaking.This approach was novel, but Eugenia Kuyda (Founder & CEO) feared it being perceived as disrespectful.Eugenia worried about tarnishing her friend's memory but believed he would embrace the idea.Her emotional drive met technical know-how.She used AI to bridge the gap between life and memory, leveraging real-time sentiment analysis and feedback loops.Today, Replika.ai has amassed well over 10M users worldwide, and continues to foster meaningful relationships through AI-driven interactions.You can read a transcript of our conversation and find more about the series at productledaipod.com
Open Source AI Will Change Everything | Mustafa Suleyman Seth Rosenberg 2024-06-24 | Closed-source, proprietary API models aren’t the only moat. Here’s Why.Previously, the gold standard was sheer size – large, complex models with billions of parameters.This approach, however, has limitations in terms of accessibility and cost.But, what if you could run a sophisticated LLM on your laptop or even your phone, with pretty much the same capabilities.Far-fetched? Not quite.Microsoft just released Phi-3.Open source + a fraction of the compute power.Does this represent a shift towards smaller, more efficient, open-source models?Time will tell.
Data Acquisition Over Parameter Count for AI? | Mustafa Suleyman Seth Rosenberg 2024-06-22 | High-Quality Data beats Parameter Count.Constantly chasing larger models is futile.Instead, the moat lies in data acquisition.As an AI founder, you need to look for existing labeled datasets or, even better, design a UI that inherently collects data from user interactions.This will help you:1. Create defensibility with high-quality, proprietary data, and2. Eliminate the need for you to depend on large-scale model providersIn this episode of Product-Led AI, I’m joined by Mustafa Suleyman of Microsoft, who clarifies why founders need to focus on data acquisition and filtering over parameter count.
AI Founders Need to Hear THIS | Mustafa Suleyman Seth Rosenberg 2024-06-22 | The blanket statement that GPT-5 will replace all human beings is a tad exaggerated.Sure, GPT-5 will be better than its predecessor, but it won't suddenly take over roles like teaching. Currently, even with the best Reinforcement Learning (RL) implementations, the results are decent but far from perfect.One hallucinatory response is all it takes to lose a user.The workaround?1. Collect high-quality data2. Filter said data3. Ensure data integrityIn this episode, Mustafa Suleyman, CEO, of Microsoft AI, stresses that delivering reliable data is HOW you win as a product-focused founder.
Will People Fall in Love With Chatbots? Seth Rosenberg 2024-06-20 | “AI Girlfriend” searches have grown 525% in the last year.1 in 5 young users is open to the idea of an AI or virtual partner.Is this indicative of a broader paradigm shift? Maybe.Replika.ai, led by Eugenia Kuyda, began with the goal of improving emotional well-being.Surprisingly, users started falling in love with their AI companions, which led to debates about limiting romantic relationships with AI.#ai #innovation #technology #entrepreneur
Preserving Relationships & Memories with Conversational AI Seth Rosenberg 2024-06-16 | Replika.ai’s journey began with a deep focus on love, life, and friendship.Launched in 2017, the project's roots trace back to a desire to build sophisticated conversational AI.The purpose was to create a memorial bot to keep conversations alive with a lost friend.This approach was novel, but Eugenia Kuyda (Founder & CEO) feared it being perceived as disrespectful.Eugenia worried about tarnishing her friend's memory but believed he would embrace the idea.Her friend's last project centered on digital avatars as memorials, blending personal loss with innovation.Eugenia’s emotional drive met technical know-how.She used AI to bridge the gap between life and memory, leveraging real-time sentiment analysis and feedback loops.Watch the full episode as Eugenia talks about how Replika.ai is redefining relationships and memories:[LINK]
Can Microsoft Co-Exist with Startups? Seth Rosenberg 2024-06-15 | Microsoft's scale opens up more doors for startups.While the tech giant focuses on horizontal fixes, it recognizes the value of custom solutions from deep domain experts.That’s where startups can innovate.Jared Spataro of Microsoft says the company aims to use its data for disruptive solutions.With over 330M DAUs for Teams, Microsoft plans to integrate startup solutions to expand its ecosystem, creating a win-win scenario.
How Replika Enabled Conversational AI Seth Rosenberg 2024-06-15 | In the early 2010s, chatbots were pretty much nonexistent.This fascinated Eugenia Kuyda, Founder & CEO of Replika.ai. She wondered why there wasn’t tech replicating our daily dialogues with computers.The way she saw it was, "Someone has to do it, it might as well be me."In 2012, breakthroughs like ImageNet and Google’s Word2Vec hinted at possibilities.Inspiration struck.The real opportunity was in building chatbots mimicking human conversation.Despite limited knowledge, Eugenia believed she was onto something.Turns out, she was right! As of 2023, Replika.ai had over 10M users worldwide.#technology #innovation #podcast #ai
Why Jeff Bezos & NVIDIA Are Backing Perplexity Seth Rosenberg 2024-06-14 | You’re an intern at OpenAI, then Deep Mind, followed by Google… Years later, you’re a Research Scientist at OpenAI, which you quit after a year to bet on yourself.Cut to 2024.You’ve built a $1B AI-based search company with investors like Nvidia, Daniel Gross, Jeff Bezos, & more.That’s Aravind Srinivas, CEO of Perplexity.ai’s story. You can read a transcript of our conversation and find more about the series at productledaipod.com
Optimizing AI Chatbots for Happiness Seth Rosenberg 2024-06-14 | AI chatbots are enhancing the human experience through empathetic interactions. Here’s how.Eugenia Kuyda, CEO of Replika.ai, took early steps to measure how these interactions affect user well-being.She focused on optimizing the ML model and user interface to favor positive reinforcement, not just substitute human interaction.The goal was to develop a dialogue system that maximizes human happiness.Eugenia and her team used post-interaction surveys and long-term metrics from psychological assessments for depression, anxiety, and social connectivity.These metrics were used to refine the reinforcement learning algorithms, ensuring that users were engaged and felt cared for.#shorts #ai #innovation #technology
Microsofts Balancing Act for Value Creation Seth Rosenberg 2024-06-14 | At the scale Microsoft is, tackling industry-specific problems is no small feat.The fix?Create tailored solutions for vertical markets across multiple use cases.As Microsoft experiments, they’re faced with a conundrum. Broad horizontal productivity, or specific vertical applications?This turning point, they find, lies in functional applications.In this conversation with Jared Spataro, he explains why the incumbent focuses on customer support.The crux of this is anchored by the foundational GPT model, which aims to increase throughput and efficiency.
Startups Need to Leverage Copilot Studio Seth Rosenberg 2024-06-13 | Copilot's core strength is its intelligent orchestration layer, which is pivotal for dissecting user prompts and refining search queries.Getting this right was a 12-month challenging journey for the team at Microsoft.This orchestration layer facilitates third parties to extend Copilot's capabilities beyond its native functionality.Startups can benefit by leveraging Copilot's intent recognition to design custom tools for specific tasks.Users can benefit from this too, by retrieving data buried within several columns of reports and spreadsheets.In this episode, Jared Spataro speaks to how Copilot Studio, a product from the house of Microsoft, is the primary means to capitalize on the orchestration layer by letting users create dedicated copilots of their own.Jared emphasizes that Copilot’s intent recognition abilities allow it to act as “a single source of truth.”
Are Chatbots Experiencing their Microsoft DOS Era? Seth Rosenberg 2024-06-12 | The best products don't just meet needs; they build connections.Shared experiences and memories make a product go beyond functionality; it speaks to the hearts and minds of users, building a lasting connection.The question on every AI founder’s mind is:How do I design a great model with memory?The answer is simpler than you think.It’s akin to the human brain's capacity to remember past conversations.Eugenia Kuyda, CEO of Replika.ai, recognized that designing memory capacity into a conversational model will be hard.Eventually, Replika.ai began allowing users to view, edit, and add to the chatbot's memory.To Eugenia’s surprise, OpenAI followed the same playbook for ChatGPT, which is the right way to go.She also talks about how chatbot interfaces remain rudimentary, and compares it to the Microsoft DOS era for AI.
Replikas Eugenia Kuyda on AI Companionship Seth Rosenberg 2024-06-11 | An AI Friend: Peculiar or Genius?Eugenia Kuyda, Founder & CEO of Replika.ai, joins me on this episode of Product-Led AI.I've known Eugenia since my days as a PM at Facebook, working on Messenger, and her story is truly inspiring.At the time, chatbots were practically non-existent.For Eugenia, Replika started from a deeply personal place after her best friend passed away.She used conversational AI to create a memorial bot.But, Replika.ai was always about more than just memorializing.It was about leveraging deep learning models for dialogue generation to replicate natural conversational experiences, which no one else seemed interested in exploring in the 2010s.Eugenia and her team still seek to answer the question: Can you improve emotional well-being, making people happier and more connected with others, and themselves?Tune in to hear Eugenia discuss the future of AI friendships, Deep Learning’s role in shaping chatbots, and optimizing chatbots for personalization moving forward.You can read a transcript of this episode on the Product-Led AI series website here: productledaipod.com
AI is Going to be Your Next Co-worker Seth Rosenberg 2024-06-10 | My guest on today’s episode is Jared Spataro, Corporate VP of AI at Microsoft, who maintains that five years ahead, copilots will become a skill at least as important as working with people if not more.For established companies, this marks a monumental shift towards AI-driven re-engineering.Jared emphasizes that as barriers to entry dissolve, this is the right time for driven and talented individuals to strike while the iron is hot.
The AI thats better than Google Seth Rosenberg 2024-06-04 | #ai #google #chatbot #innovation #technologyTraditional LLMs & search engines (even Google) are deeply flawed.While LLMs excel at generating fluent text, their accuracy isn't guaranteed.They prioritize coherence within a sequence over factual correctness, which makes LLMs (even ChatGPT) prone to hallucinate more.As for traditional search engines, they rely on clicks for revenue.This leads to a mix of high-quality sources alongside less reliable ones in search results.Perplexity.ai’s Founder & CEO, Aravind Srinivas aimed to mitigate these issues by combining search index capabilities with LLM summarization.Aravind views Perplexity as an answer engine that provides the answer to a query directly and the source alongside.After launch, Aravind and his team realized they were a threat to Google’s core business: search.Little did they know, this was just the beginning.
AI Chatbots - The next BIG THING in ad revenue Seth Rosenberg 2024-06-01 | #ai #google #chatbot #innovation #technologyGoogle generates almost $1B/day in ad revenue…but that’s all about to change.Perplexity.ai’s co-founder & CEO Aravind Srinivas is confident that AI chatbots will create new avenues for content discovery and engagement.If chatbots lead the way, this will require educating advertisers on bidding strategies and ad placements.Although AI-driven ad models might operate with lower margins, startups could thrive on positive margins, even if they are lower than the incumbents.Who do you think will win this race, startups or Big Tech?
Why Perplexity has an Edge Over Google Seth Rosenberg 2024-05-30 | #AI #google #perplexity #shorts #podcast Perplexity.ai’s search index is a bone of contention for Big Tech, including Google, which has an existing index but struggles with:A) Curating it to specialized applications like AI chatbotsB) Inadequate indexing infrastructure leading to strained efforts in crawling and rankingPerplexity, however, is winning with its search indexing capabilities because of:A) Focused attempts to scrape relevant informationB) Tailored indexing strategies for domain-specific applications, like AI chatbotsC) Leveraging user interactions to improve the effectiveness of search indexingIn this episode of Product-Led AI, Perplexity co-founder & CEO Aravind Srinivas also explains how search indexing can be a moat for the company for years to come.
Perplexity CEO Aravind Srinivas on Disrupting Search with AI Seth Rosenberg 2024-05-29 | You’re an intern at OpenAI, then Deep Mind, followed by Google… Years later, you’re a Research Scientist at OpenAI, which you quit after a year to bet on yourself.Cut to 2024.You’ve built a $1B AI-based search company with investors like Nvidia, Daniel Gross, Jeff Bezos, & more.That’s Aravind Srinivas, CEO of Perplexity.ai’s story. In this episode of Product-Led AI, Aravind and I discuss a host of topics– Perplexity’s moat as the AI-powered "answer engine" that is disrupting search, how this could impact the ad-based business model of the sector, why startups have an edge on incumbents, and more.You can read a transcript of our conversation and find more about the series at productledaipod.com0:00 - Introduction to AI applications1:06 - Aravind's background and journey to founding Perplexity2:38 - Evolution of Perplexity's concept and funding4:08 - Transition from research to end-user product5:30 - Solving the "hallucination" problem7:45 - Creating a focused product on web search10:33 - Summarization for faster answers11:48 - Characteristics of Pro subscribers13:00 - Potential paths for Perplexity's growth15:10 - Data security concerns in workplace browsers18:01 - Categorizing models as taker, shaper, maker20:23 - The challenge of indexing the whole web21:38 - Long-term commitment to indexing and crawling24:01 - Importance of routing to appropriate models25:24 - Ensuring new websites can become highly successful27:10 - Future of citation economy and web monetization
Microsofts Jared Spataro on Rewiring How We Work Seth Rosenberg 2024-05-23 | Jared Spataro has been leading work modernization efforts for Microsoft for more than 15 years, helping to launch and scale products that enterprise organizations use every day while keeping up with the latest tech shifts. AI has made every part of his job move faster. Now, a little more than a year after the release of Microsoft's Copilot for 365, Jared talks about Microsoft’s AI strategy; the technical architecture of the AI tool works alongside people; what they learned in its first year of deployment and how it is impacting enterprise organization; his vision of the future working with AI agents and copilots; and where he sees opportunities for startups.You can find more about this episode and the entire Product-Led AI series at productledaipod.com
Adept CEO David Luan on Upleveling Human Work Seth Rosenberg 2024-05-14 | Adept CEO and co-founder David Luan has been riding the LLM wave since the early days of the modern AI boom: an early OpenAI employee, he led the company’s engineering team before serving as the tech lead for Google Brain’s work on LLMs. He co-founded Adept in 2022 to develop multimodal agents built to work alongside humans in any profession. Believing computers should work more like humans and not the other way around, Adept has stood out for its human-centric approach to AGI. In this episode, David talks about the moment he realized building a new product (and a startup) was a faster track to AGI; breaks down the frameworks for Adept; and shares his view of a future where AI functions almost invisibly below all software we use. Check out the whole series at productledaipod.com
Reid Hoffman on AI-Powered Networks Seth Rosenberg 2024-05-07 | How could AI be the catalyst for the next wave of iconic networks? It helps to remember that regardless of the platform shift that gives rise to a new breed of networks, there are various components that must come together for them to actually work. From LinkedIn and Meta to Airbnb and TikTok, the fundamentals are the same: clear value proposition, compounding growth and engagement loops, and a strategy to leverage the new tech to drive it forward. Greylock partner, AI visionary, and renowned network-builder Reid Hoffman shares his thoughts on AI’s potential to fuel the next generation. Additionally, Hoffman weighs in on the broad opportunities for AI startups building in the application layer.You listen to the audio version of this interview on all major podcast platforms, and learn more about this episode on the series website productledaipod.com
Ramp CEO Eric Glyman on Building the AI Platform for Finance Teams Seth Rosenberg 2024-04-30 | Fintech company Ramp has always been obsessed with time. The one-stop-shop financial operations platform was purpose-built to make businesses more efficient so they can focus on the work that really matters – and create more wealth. Since launching in 2020, the company has saved businesses 10 million hours of human labor and more than $1 billion in spend. AI could make that impact even higher, says Ramp CEO Eric Glyman. The company has gone all in on AI both internally and externally.In this episode, Eric talks about fintech’s natural fit for AI; how Ramp has thoughtfully incorporated AI capabilities into the platform in recent years; and shares his outlook for the day when all knowledge work is enhanced by AI, giving people the time they need to focus on creating and building. Check out the whole series at productledaipod.com
