Is AI Revenue Inflated? | The Dejima Mafia

Is AI Revenue Inflated? | The Dejima Mafia

Babak Hodjat has been building AI agents since 1998. He said that in front of a room in San Francisco last week, sitting next to his brother Siamak for the first time in 25 years. His argument is narrower than the bubble talk, and more useful. Enterprises are not buying AI because a need has been identified. They are buying because the business side expects productivity in return. The engineering that would actually deliver that productivity has not been done yet. Until it is, the revenue is running ahead of the value. He also thinks most of the layoffs attributed to AI were cuts companies could already afford. Babak co-founded Dejima in 1998. The patent behind it, "Adaptive interaction using an adaptive agent-oriented software architecture," carries a June 15 1998 priority date and became foundational to Apple's Siri. Sybase acquired the company in 2004. He went on to co-found Sentient Technologies, applying evolutionary algorithms to trading, and now leads AI at Cognizant. Siamak Hodjat was at Dejima from the start and stayed in the assistant world through Siri, Viv Labs and Gameplanner alongside Adam Cheyer. Recorded live in San Francisco on July 25, 2026, in front of a room of founders and researchers. Our Partner : www.humanagency.com [http://www.humanagency.com] News letter : https://secondnewsletter.substack.com [https://secondnewsletter.substack.com] Babak Hodjat : https://www.linkedin.com/in/babakhodjat/ [https://www.linkedin.com/in/babakhodjat/] Siamak Hodjat: https://www.linkedin.com/in/siamak-hodjat-98669/ [https://www.linkedin.com/in/siamak-hodjat-98669/] Caroline Hoogland: https://www.linkedin.com/in/caroline-hoogland-a092b25a/ [https://www.linkedin.com/in/caroline-hoogland-a092b25a/] Ardalan Javadi: https://www.linkedin.com/in/ardalanjam1369/ [https://www.linkedin.com/in/ardalanjam1369/] Farzam Hejazi: https://www.linkedin.com/in/farzam-hejazi-5b608081/ [https://www.linkedin.com/in/farzam-hejazi-5b608081/] About Us : Linkedin : / secondpodcast Twitter : /thesecondpod Instagram : / thesecondpod #Secondpod #Second_Pod #پادکست ـ دوم #secondpod #پادکست

Ramin Hassani I Co-Founder & CEO of Liquid AI

Ramin Hasani is the co-founder and CEO of Liquid AI, and he has done what almost no one in AI has: built a real alternative to the transformer, the architecture behind ChatGPT and nearly every model you touch. Liquid AI raised a $250M Series A led by AMD, and by his own account now sits around $4.5 billion. The core idea did not come from a bigger data center. It came from the 302-neuron brain of a 2mm worm. We also get into the part most founders never put on a slide. While the whole field was stripping complexity out of neural networks to make them scale, Ramin spent seven years adding it back in, chasing an equation about how two neurons talk that had sat unsolved since 1907. He cracked it on a chalkboard at MIT, one now headed to the MIT history museum. Ramin was born in Iran and left for Politecnico di Milano, then a PhD at TU Wien in Vienna, where modeling the brain of a worm became Liquid Neural Networks. From there: MIT, CSAIL, a 2022 Nature paper, and in March 2023 spinning the research into Liquid AI alongside Daniela Rus, Mathias Lechner, and Alexander Amini. Our News Letter : https://substack.com/@seondnewsletter [https://substack.com/@seondnewsletter] Ramin Hassani Linkedin : https://www.linkedin.com/in/alikashani [https://www.linkedin.com/in/alikashani] Ardalan Javadi : https://www.linkedin.com/in/ardalanjam1369/ [https://www.linkedin.com/in/ardalanjam1369/] Farzam Hejazi: https://www.linkedin.com/in/farzam-hejazi-5b608081/ [https://www.linkedin.com/in/farzam-hejazi-5b608081/] 00:00:00 Introduction  00:03::24 Why Noise Is Actually a Resource for Computation  00:06:01 Where the Whole Worm Idea Came From  00:6:56 Why Ramin Chose the C. elegans Worm as the Basis of His Research  00:10:19 How Ramin Found His First Co Founder  00:18:08 The Equation Ramin and His Team Solved That Had Been Unsolved Since 1907 00:19:06 The First Mathematical Attempt in 1953 to Show Neuron Relationships 00:19:41 What Is the Problem With Differential Equation Based Machine Learning? 00:21:29 The Aha Moment Behind Liquid AI  00:22:16 The Birth of Liquid AI  00:22:46 How Sergey Levine Inspired the Name  00:24:09 How Liquid AI's Architecture Differs From Transformers  00:25:09 The Relationship Between Bias and Algorithm Scalability  00:28:05 Why Liquid Neural Networks Have an Efficiency Advantage at Scale  00:29:41 Do Scaling Laws Apply to Liquid Models?  00:31:26 Attention Is Not the Only Thing You Can Scale  00:34:08 How the Liquid Search System Works  00:34:53 Why Liquid AI Has the Lowest Memory Usage and Latency and Why It Is Hybrid  00:35:00 The Best Foundation Model for CPUs  00: 37:30 How to Close Enterprise Customers in the AI Era  00:38:29 What Vanguard Taught Ramin About Financial Strategy  00:41:58 125 Enterprises Are Now Using the Liquid AI Model  00:42:56 How to Design a Revenue Model for Enterprise in the AI Era00:43:19 How Shopify Uses Liquid AI 00:44:00 How Mercedes Uses Liquid AI With a 600 MB Model  00:45:51 The Quality of $1 of Liquid AI Revenue Compared to $1 of Anthropic  00:47:13 Why Liquid Is Going After Arm's Business Model  00:52:53 The Story Behind Another Iranian Founded AI DNA Startup That Raised $200M in Seed  00:55:47 The Story Behind Liquid AI's $290M Fundraise  00:59:30 How to Close a Seed Round  00:1:00:17 The Mastermind Behind the Fundraising  00:1:02:57 The Story Behind AMD Leading the Round  1:04:56 Why Revenue Matters So Much for Fundraising  1:05:22 Liquid AI Is Now at a $4.5 Billion Valuation  1:05:51 Ramin's Insight on Building a World Class Network  1:07:32 Ramin's Takeaways From Meeting Jamie Dimon  1:08:31 Jamie Dimon's Pitch for Liquid AI  1:08:51 How to Build an Effective Short Narrative for Your Pitch  1:10:31 Leading by Example  1:11:07 What Is the Best Definition of Intelligence?  1:14:32 Silicon Valley Is a Gossip Town  1:16:22 Why an Exponential View Matters and How Sam Altman Does It References: Check our News Letter : https://substack.com/@seondnewsletter?r=2v0lej&utm_campaign=profile&utm_medium=profile-page [https://substack.com/@seondnewsletter?r=2v0lej&utm_campaign=profile&utm_medium=profile-page] About Us : Linkedin : https://www.linkedin.com/company/secondpodcast/ [https://www.linkedin.com/company/secondpodcast/] Twitter : /thesecondpod Instagram : https://www.instagram.com/thesecondpod/ #Secondpod #Second_Pod #پادکست ـ دوم #secondpod #پادکست

Young AI Researcher Disrupting the Entire Industry - Ali Behrouz

Ali Behrouz is an AI researcher and the author of two of the most talked about papers in machine learning right now, Titans and Nested Learning. He is a PhD student at Cornell University and a researcher at Google, and his work takes direct aim at one of the field's biggest questions: what comes after the transformer. His paper Nested Learning, which Google's Jeff Dean called a possible paradigm shift, rethinks the whole deep learning stack. In this conversation we get into the back story behind Titans and Nested Learning, how Ali thinks about transformers and where they fall short, the idea that all of deep learning is really associative memory, and his newer work on giving models a kind of "sleep" phase to consolidate what they learn. One of the most insightful conversations we have had on where AI is heading. Ali was named as the best AI researcher in 2025 by Second Group: Second Prize Awards 2025, Celebrating the Best Iranian AI Scientists, Roboticist, CEOs & Startups [https://www.youtube.com/watch?v=9kLQXX76rEI&t=1732s] Ali Behrouz : Ali BehrouzGitHubhttps://abehrouz.github.io [https://abehrouz.github.io/] Ardalan Javadi: https://www.linkedin.com/in/ardalanjam1369/ [https://www.linkedin.com/in/ardalanjam1369/] Farzam Hejazi: https://www.linkedin.com/in/farzam-hejazi-5b608081/ [https://www.linkedin.com/in/farzam-hejazi-5b608081/] Subscribe to our Newsletter : https://substack.com/@seondnewsletter [https://substack.com/@seondnewsletter]   00:00 Introduction  03:26 What is Intelligence?  05:51 Academic Journey 11:57 From Blockchain to Brain Networks 17:21 The Allure of Computational Social Science 21:02 The Genesis of TITAN 24:39 Human Memory Systems vs. AI Caching 32:36 The Disconnect: Solving Anterograde Amnesia in LLMs 41:40 Mechanics of TITAN 47:37 The Viral Aftermath: Dealing with Hype and Haters 52:09 Why Transformers Dominate & The Rebranding Epidemic 57:24 Scaling Laws and the Limitations of Pure Compression 64:05 Anthropic vs. OpenAI 66:06 Architectural Frontiers: Loop Transformers & Mixture of Depths 69:01 Moving Beyond TITAN 77:51 Nested Learning: Emulating Brain Frequencies 81:44 Memory Distillation and AI Sleep Cycles 83:24 Career Advice: Scientific Progress as "Old Ideas + New Noise"   References https://www.seltzer.com/margo/ [https://www.seltzer.com/margo/] https://en.wikipedia.org/wiki/Large-scale_brain_network [https://en.wikipedia.org/wiki/Large-scale_brain_network] https://www.geeksforgeeks.org/machine-learning/introduction-to-recurrent-neural-network/ [https://www.geeksforgeeks.org/machine-learning/introduction-to-recurrent-neural-network/] https://en.wikipedia.org/wiki/Mamba_(deep_learning_architecture) [https://en.wikipedia.org/wiki/Mamba_(deep_learning_architecture)]   About Us :  Linkedin : https://www.linkedin.com/company/secondpodcast/ Twitter : /thesecondpod    Instagram : https://www.instagram.com/thesecondpod/ #Secondpod #Second_Pod #پادکست ـ دوم #secondpod #پادکست

Ali Kashani, Serve Robotics

Ali Kashani is the co-founder and CEO of Serve Robotics, and he has done what almost no one in Robotics  has: put robots on real streets in dozens of cities and made people love them. He is the youngest Iranian-born CEO on the Nasdaq, the youngest CEO of any robotics company on the exchange, growing revenue around 600 percent year over year and scaling from 100 robots to thousands in a single year. We also get into the part most founders never put on a slide. Between spinning out of Uber and ringing the Nasdaq bell, Serve hit its last payroll ten times. Two of those times Ali made payroll out of his own mortgage.  Ali was born in Iran and left at 17. He did his robotics PhD at the University of British Columbia on fusing LiDAR and cameras, started an energy hardware company that nearly died and then sold  I co-hosted this one with a friend of the pod Brendan, Managing Director of AI and Partner at Human Agency [https://www.humanagency.com/].   Our News Letter : https://substack.com/@seondnewsletter [https://substack.com/@seondnewsletter] Ali Kashani Linkedin : https://www.linkedin.com/in/alikashani [https://www.linkedin.com/in/alikashani] Ardalan Javadi : https://www.linkedin.com/in/ardalanjam1369/ [https://www.linkedin.com/in/ardalanjam1369/]   00:00:00 The youngest Iranian-born CEO on the Nasdaq    00:01:19 Why a public company beats a private one  00:02:47 The path to Serve: a robotics PhD and LiDAR 00:04:14 Fired, broke, and one step at a time  00:07:20 The dead company that came back  00:08:31 DARPA 2005 and the contrarian bet on LiDAR  00:12:45 Designing a robot people would forgive  00:16:04 Principle before perception  00:18:45 Building the ecosystem: NVIDIA, Magna, Uber  00:23:56 Dots versus lines, and "Exit Path"  00:26:24 Turning customer concentration into a feature  00:29:38 The real unit economics of a delivery  00:31:48  Picking cities and rolling out robots  00:33:17  The case for optimism on AI and jobs  00:37:49  Sidewalks versus streets  00:41:18  The IPO and the numbers  ‘00:42:23  Ten last payrolls and the Kashani loan  00:46:10  Two days before Christmas: the call to Jensen  00:48:50  From bankruptcy lawyers to the bell in three months  00:49:11  The coming valley and the humanoid hype  00:50:50 Where the value is in physical AI  00:53:24 Why narrow robots beat humanoids, for now  00:55:00 Don't follow trends  00:57:44 Uncertainty is the job  01:00:59  Startups as entropy reduction engines  01:02:31 Touraj, Noosheen & Pejman 01:04:39  His hope for Iran References :    https://investors.generac.com/news-releases/news-release-details/generac-announces-acquisition-neurio-technology-inc [https://investors.generac.com/news-releases/news-release-details/generac-announces-acquisition-neurio-technology-inc] https://en.wikipedia.org/wiki/DARPA_Grand_Challenge [https://en.wikipedia.org/wiki/DARPA_Grand_Challenge] https://bothsidesofthetable.com/invest-in-lines-not-dots-611f36491d73 [https://bothsidesofthetable.com/invest-in-lines-not-dots-611f36491d73] https://www.therobotreport.com/four-ways-to-make-ai-products-that-people-will-love/ [https://www.therobotreport.com/four-ways-to-make-ai-products-that-people-will-love/] About Us :  Linkedin : https://www.linkedin.com/company/secondpodcast/ [https://www.linkedin.com/company/secondpodcast/?viewAsMember=true] Twitter : /thesecondpod    Instagram : https://www.instagram.com/thesecondpod/ [https://www.instagram.com/thesecondpod/] #Secondpod #Second_Pod #پادکست ـ دوم #secondpod #پادکست

$3.2B in 10 Years | Martin Basiri

Martin Basiri built ApplyBoard into a $3.2B company in 10 years, helping over 2 million students study abroad. He is among the very few Iranian immigrants to build a multibillion dollar company in North America, and ApplyBoard alone now contributes 0.6% of Canada's GDP. In this first time Martin shares his full playbook, before his first $3M raise, his vision for the future of immigration, and what he is building next with Passage, a new financial aid platform reimagining how students fund their education. Martin Basiri Linkedin : https://www.linkedin.com/in/martinbasiri/ [https://www.linkedin.com/in/martinbasiri/] More about Passage : https://www.linkedin.com/posts/martinbasiri_passage-is-3-years-old-we-started-the-share-7445530978036658176-fnn4?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAABcWhnUBc4VY_Jeg2Zkx0n3jQpTTTh3xu30 [https://www.linkedin.com/posts/martinbasiri_passage-is-3-years-old-we-started-the-share-7445530978036658176-fnn4?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAABcWhnUBc4VY_Jeg2Zkx0n3jQpTTTh3xu30] Our Newsletter : https://secondnewsletter.substack.com/ [https://secondnewsletter.substack.com/] Ardalan Javadi : https://www.linkedin.com/in/ardalanjam1369/ [https://www.linkedin.com/in/ardalanjam1369/] 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒  0:00 0.6% of Canada's GDP  0:39 Martin's Introduction  2:02 The system design idea behind ApplyBoard  7:37 Did your short experience as an employee in Ohio make it clear you couldn't work for others?  9:09 Martin's philosophy on taking action 1 1:09 Martin's impact on Canada's economy  13:01 ApplyBoard is the largest study abroad company globally  13:44 The global problem in higher education  14:51 ApplyBoard as the greatest impact entrepreneurship  17:00 Passage's financial aid system  18:22 How did Martin discover the higher education problem?  28:22 Contrarian and right  29:20 Martin's insight on growth stage capital  33:41 ApplyBoard's revenue growth stats  34:52 300 no's for the first $3M  35:18 Plug and Play  39:25 Martin is the only Iranian in the past  20 years to build a $3.2 billion company as an immigrant  40:37 Martin's take on AI's long term impact on higher education  48:39 Martin's take on the future of immigration  50:00 Fertility rate  57:07 Martin's prediction on the future of study abroad and why Passage matters  1:02:00 Martin's unique insight on immigration  1:04:00 Canada's population decline this year  1:05:22 Canada's productivity decline 1:07:08 The median income of immigrants from Canada to the US is $142,000 1:08:14 The rising cost of tuition in higher education 1:09:10 What is Passage? 1:14:16 Which country has had the best immigration policy in recent years? 1:18:19 Martin Basiri's 5 toolkits for entrepreneurs