Tom Masterson, COO of Deep Genomics, on AI drug discovery where precision beats creativity. He explains how the company's siRNA model picked an approved cholesterol drug out of 3,500 candidates for a target it had never seen, why it designs toxic molecules on purpose, and how a COO runs a company where the lab and the software move on different clocks.

In this episode, Tom and Michael discuss:

• The difference between life sciences and healthcare, and where a tech bio company fits
• Brendan Frey, Geoffrey Hinton's lab, and a thesis that was right but early
• Retiring 40 task-specific models when a foundation model beat them
• Lab in the loop, and why the best training data includes molecules designed to be toxic
• The siRNA result: a target the model had never seen, and nine picks out of 3,500
• Why drug developers think in 10 years and machine learning teams think in an afternoon
• From chief of staff to COO, and getting scientists to adopt process
• Running waterfall and agile side by side, on a quarterly board and advisor rhythm
• The N-of-one problem: a hero to the world, or a hero to your shareholders

About Tom:
Tom Masterson is chief operating officer of Deep Genomics, the Toronto company building biological foundation models for RNA drug discovery. He joined in business development and strategy, then served as chief of staff to founder Brendan Frey and to CEO Brian O'Callaghan before becoming COO. He partners with the CEO on board and investor relations, owns the company's project management and executive functions, and is the person who turns Deep Genomics' science into its business story. Before Deep Genomics he was the first employee at Microbiome Insights in Vancouver, worked in contract research and healthcare design strategy, and founded Caregiver Support, a startup that helped family caregivers through medical crises. He started his career selling dental lasers at a medical device company. He holds a BSc in cell biology and genetics from the University of British Columbia and an MBA from Harvard Business School.

Resources mentioned:

• Deep Genomics: https://www.deepgenomics.com
• GenomeKit (open source): https://github.com/deepgenomics/GenomeKit
• Alnylam Pharmaceuticals: https://www.alnylam.com
• Moderna: https://www.modernatx.com
• Merck: https://www.merck.com
• Broad Institute: https://www.broadinstitute.org

Connect with Tom:

• LinkedIn: https://www.linkedin.com/in/mastersonthomas/

Connect with Michael:

• LinkedIn: https://www.linkedin.com/in/michael-koenig514
• Building Helm: https://helmapp.ai

Subscribe to Between Two COOs:

• Apple Podcasts: https://podcasts.apple.com/us/podcast/between-two-coos-with-michael-koenig/id1596195041
• Spotify: https://open.spotify.com/show/4OQY2oizo3rWUlwhmtdpSi
• Newsletter: https://betweentwocoos.com

Paul Kedrosky of SK Ventures says AI capex is driving up to 70 percent of US GDP growth, and most of it is now borrowed. He walks through the money map, the 2029 maturity wall, why AI tokens are a deflationary commodity, and three ways the AI bubble reaches companies that never touch a data center.

In this episode, Paul and Michael discuss:

• Why AI capex now accounts for 30 to 70 percent of US GDP growth, and the historical moments that looked the same
• The money map: how the buildout moved from hyperscaler cash flow to more than 60 percent external financing
• The 2029 maturity wall, and why it rhymes with the mortgage resets of the financial crisis
• Tokens as the first hyper-deflationary industrial commodity, and the growth math that follows
• Why developers are a misleading signal for how the rest of the economy will use AI
• Circular financing: when one company is customer, supplier, investor, and lender at once
• Whether anyone is safe building on top of the frontier models, and what quant funds do instead
• Three things operators should do now: guard your signal, vet your vendors, and stop extrapolating from coders

About Paul:
Paul Kedrosky is co-founder and managing partner of SK Ventures and a fellow at MIT’s Initiative on the Digital Economy, where his research focuses on AI and the future of work. A former Wall Street technology equity analyst, he consults with major asset managers, is a regular contributor at CNBC and Bloomberg, and co-hosts a weekly show with former Twitter CEO Dick Costolo.

Resources mentioned:

• Paul Kedrosky: https://paulkedrosky.com
• MIT Initiative on the Digital Economy: https://ide.mit.edu
• National Bureau of Economic Research: https://www.nber.org
• CoreWeave: https://www.coreweave.com
• Hudson River Trading: https://www.hudsonrivertrading.com
• Jane Street: https://www.janestreet.com
• Stratechery by Ben Thompson: https://stratechery.com

Connect with Paul:

• LinkedIn: https://www.linkedin.com/in/paul-kedrosky/
• Newsletter: https://paulkedrosky.com

Connect with Michael:

• LinkedIn: https://www.linkedin.com/in/michael-koenig514
• Building Helm: https://helmapp.ai

Subscribe to Between Two COOs:

• Apple Podcasts: https://podcasts.apple.com/us/podcast/between-two-coos-with-michael-koenig/id1596195041
• Spotify: https://open.spotify.com/show/4OQY2oizo3rWUlwhmtdpSi
• Newsletter: https://betweentwocoos.com

Lior Eldan, co-founder and COO of Moburst, has reinvented what his agency sells three times, from app store optimization to AI products. He explains why replacing people with AI is backwards, why managers have to do the work, and why automating a broken process just gets the AI blamed.

In this episode, Lior and Michael discuss:

• Why Moburst has moved off its winning product three times since 2013, and what each move broke
• Build vs. buy for services: why an acquisition brings a founder, and what the day after actually looks like
• Managers do the work: the rule that keeps middle management from becoming a layer of measurers
• Why replacing people with AI is backwards, and what Moburst now requires of every junior hire
• Forward-deployed: what an embedded team found inside a mortgage company's operations
• Answer engine optimization and why an LLM reads your whole digital footprint, not your website
• Fix the process before you automate it, and the cadence underneath a company that keeps reinventing itself

About Lior:
Lior Eldan is co-founder and COO of Moburst, the digital marketing agency he started in 2013 with Gilad Bechar after the two met as HR officers in the Israeli army. Moburst began in app store optimization, grew through acquisitions including Uproar and Kitcaster, and now sells AI products alongside its services. Adweek has named it one of the fastest-growing agencies in the world.

Resources mentioned:

• Moburst: https://www.moburst.com
• Kitcaster: https://kitcaster.com
• HubSpot: https://www.hubspot.com
• Supercell: https://supercell.com
• Claude Code: https://claude.com/product/claude-code
• Tucows: https://tucows.com

Connect with Lior:

• LinkedIn: https://www.linkedin.com/in/lioreldan/

Connect with Michael:

• LinkedIn: https://www.linkedin.com/in/michael-koenig514
• Building Helm: https://helmapp.ai

Subscribe to Between Two COOs:

• Apple Podcasts: https://podcasts.apple.com/us/podcast/between-two-coos-with-michael-koenig/id1596195041
• Spotify: https://open.spotify.com/show/4OQY2oizo3rWUlwhmtdpSi
• Newsletter: https://betweentwocoos.com