Welcome, humans.

Did you know that the average new drug takes over a decade and roughly $6B to reach the market? Or that ~ 90% of compounds that enter clinical trials fail?

That's the hard wall that the entire pharma industry has been hitting for decades.

Well, you probably do know about AlphaFold , the Google DeepMind AI that solved a 50-year-old protein folding problem and won its creators the 2024 Nobel Prize in Chemistry .

What you might not know: the team behind AlphaFold 3 is now designing real drugs, and they just published a technical report showing their new system more than doubles AlphaFold 3's accuracy on the hardest part of drug design.

That team is Isomorphic Labs , a DeepMind spinout led by Demis Hassabis. And in our latest podcast episode , we sit down with Becky Paul (who leads medicinal drug design) and Michael Schaarschmidt (who leads foundational AI research) at Isomorphic Labs to unpack what AlphaFold actually unlocked, why drug discovery is still brutally hard, how AI drug discovery works from a process standpoint, and where (as well as how) their AI is starting to design molecules that human chemists couldn't.

Our favorite moments:

• ( 2:13 ) The brutal math of drug discovery: A decade of work. ~$6B per drug. 90% of clinical trial entrants fail. Becky lays out exactly why pharma needs something disruptive.

• ( 7:27 ) AI just replaced a PhD's worth of work: Their structure prediction model now matches X-ray crystal structures so well, the team is starting to skip the (slow, expensive) experimental step entirely.

• ( 12:19 ) Why AlphaFold 3 took 2 years for anyone else to copy: Even with the full paper, open code, and inference algorithms published, no other lab could fully reproduce it. Michael on the modeling details that don't fit in a paper.

• ( 15:11 ) "I've never worked somewhere where the pace of innovation is so fast": Becky on what it's like when the ML team next door drops a new foundation model that suddenly unblocks a project that was stuck.

• ( 19:08 ) Beyond binding, drugs that delete proteins: The molecular glue revolution, where new drugs don't just block a target, they tag it for destruction inside your cells.

• ( 26:37 ) Becky’s pie-in-the-sky dream: One design cycle. A few hundred molecules. A drug candidate at the end. What "AI-first drug discovery" actually looks like at the limit.

• ( 36:36 ) "How did it know that?": The most exciting moments aren't when the model is right, they're when the model makes a leap the team can't explain.

• ( 40:13 ) The "undruggable" protein, drugged: KRAS was labeled untouchable for decades. Recent KRAS drugs are now doubling survival in pancreatic cancer . Becky and Michael's bet: AI shrinks the next "undruggable" target from a multi-decade slog to something tractable.

Why watch this? Most "AI in healthcare" news to date has been vague optimism or more administrative than groundbreaking. This is the opposite. Two practitioners walking through exactly which steps of drug design are now AI-first, which still need wet-lab work, and where the next breakthroughs are coming from.

If you've ever wondered what AlphaFold actually means for patients (not just scientists), or why Demis keeps saying biology is the ultimate AI application, plus how the team at Isomorphic approaches the incredibly complicated task in front of them, then this is the clearest 50 minutes you'll find on it.

Watch and/or Listen now: YouTube | Spotify

P.S. At (7:27) Becky describes the moment a chemist has to stop validating the model and start trusting it. That's a quietly wild thing for a scientist to say out loud….

Dive deeper with these resources:

• Technical report blog (the easiest read, with all the visuals).

• Full technical report (PDF) (for the chemists and ML folks).

• Isomorphic Labs (they're hiring across Boston, London, and Lausanne !).

Real quick: Want to see your AI-adjacent product or service show up right here, below these podcast promos? Click the button below to advertise to our 700K+ readers!

THIS EPISODE WAS BROUGHT TO YOU BY…

We use ClickUp to publish 2,000+ pieces a month.

Across 18 spaces and 100 contributors, ClickUp is the operating system that holds our parent company (TechnologyAdvice)'s work all together. ClickUp's AI layer ( ClickUp Brain ) plugs into your tasks, docs, Slack, Gmail, and Salesforce; you pick the underlying model (Claude, GPT, or Gemini). Active users save 1.1 days weekly and see 3x faster task completion .

Oh, and did we mention the Free Forever plan has unlimited tasks and users?! Check it out!

→ Try ClickUp free

🔴 LIVE TOMORROW: The AI Starter Kit, What to Try… and What to Ignore

This week, Grant and Corey are going to cut through the AI noise for anyone who's still on the sidelines. New to AI, or know someone who is? This is the session to send them.

We'll cover:

• The best first steps for AI beginners

• What tools and features are actually worth trying right now

• What you can safely ignore for the moment

• Simple ways to get better answers from any AI tool

• How to troubleshoot when AI gives you something unhelpful

Choose your favorite platform to watch live: Watch on YouTube | Join on LinkedIn

🎙️ In Case You Missed It…

Four recent interviews you’ll definitely want to check out (pick whatever looks interesting to you and dive in!):

1. Interested in what's missing before we hit AGI? Watch: This Company Mapped the Entire World in 3D. Here's Why.

TL;DW: Peter Wilczynski, CPO at Vantor (formerly Maxar), built a 3D model of the entire Earth at 50cm resolution and made it machine-readable. He argues spatial intelligence is the gap nobody's talking about in AI, and probably the missing piece before agents can actually operate in the physical world.

Why you should watch: If you've ever wondered why AI can write code and solve math olympiad problems but still can't reliably tell a drone where to go, this one answers it. Also, there's a wild bit about how the physical world becomes the new navigation layer for AI agents.

• YouTube: Watch Here

• Spotify: Listen Here

• Apple Podcasts: Listen Here

2. Curious how good AI music tools have actually gotten? Watch: This AI Just Made Our Podcast Theme Song

TL;DW: Corey sits down with Kendall Rankin , who left LinkedIn in 2024 to join Producer AI when it was a startup (advised by The Chainsmokers, no less). Google acquired the team in February 2026, and Kendall is now on the Flow Music team inside Google Labs. On the episode, they generate a garage rock song from a single sentence, build a custom synth in the "Spaces" feature, and walk through SynthID watermarking and one-shot music videos.

Why you should watch: Most AI music demos hand you a polished finished song and skip the part where things go sideways. This episode is the part where things go sideways. First pass fumbles, Corey asks for "more fuzz," second pass actually lands. That iteration loop is the whole story for anyone trying to figure out if these tools are actually usable.

• YouTube: Watch Here

• Spotify: Listen Here

• Apple Podcasts: Listen Here

3. Want agents that actually work on real tasks? Watch: Inside the Secret Labs Where AI Learns to Work

TL;DW: Nick Heiner, VP of Product at Surge AI (a $1.2B-revenue company built without VC money), reveals why even GPT-5, Claude, and Gemini still fail about 40% of real workplace tasks, what makes a good RL environment, and his bold prediction of a $1B company with one human employee by 2030.

Why you should watch: If you're trying to get AI agents to actually finish real work (and not just demo well), this is the missing piece on why they keep falling short.

• YouTube: Watch Here

• Spotify: Listen Here

• Apple Podcasts: Listen Here

4. Interested in where AI meets actual factories? Watch: 80% of US Factories Have Zero Robots. Google Wants to Fix That.

TL;DW: Intrinsic (a Google Alphabet company) CTO Brian Gerkey co-created ROS , the open-source software running more than 1 million robots including NASA's. He explains why most US factories still have zero automation, and how generative AI is about to change that.

Why you should watch: If you think "AI and robotics" is still science fiction for your industry, this interview resets the timeline. The factory floor is where the next AI wave actually lands.

• YouTube: Watch Here

• Spotify: Listen Here

• Apple Podcasts: Listen Here

One more before you go:

If today's episode got you thinking about the AlphaFold side of DeepMind, our recent interview with Ioannis Antonoglou (one of the AlphaGo co-creators) is the natural next watch.

Different DeepMind alum, completely different bet on what frontier AI looks like next. This is very relevant now that the US government will start vetting AI models before they’re releases, or potentially outright banning AI models from certain countries.

Last thing: And if you haven’t subscribed yet, please do! Click the image below to go to our channel and hit “subscribe” to get notified right when new videos go live.

We have a goal to hit 50K subscribers by the end of the year (if not 100K), and we’re only ~30K away! If you like learning about AI, and already watch some of our videos, do us a favor and click here to subscribe today.

Stay curious,

The Neuron Team

P.P.S: Love the newsletter, but don’t want to receive these podcast announcement emails? Don’t unsubscribe — adjust your preferences to opt out of them here instead .