Artificial Intelligence

[brobro777]

https://www.bbc.com/news/articles/cwy7vrd8k4eo
 

see man this is what I'm talking about. gates, Altman, the Economist and now the BBC all say AI is a bubble? No bro, during the top of the bubble, you're supposed to get comments from prominent  guys like Bernanke saying soft landing in 2007: https://www.ft.com/content/1262b8fe-bc44-11db-9cbc-0000779e2340

 

 

[investmd]

@gfp - I agree that Insilico and Isomorphic Labs are doing cool things and likely building good businesses. 

 

The 3 things that companies like Insilico and Isomorphic Labs do is: 

1. Improve speed to market
2. Decrease cost of drug development
3. Improve success rate of drug development

As you point out these companies are partnered with Big Pharma. My point is the above 3 value propositions improve margins at Big Pharma. 

 

In terms of a basket approach, the following 4 are top 10 Big Pharma by market cap with P/E of <20 & dividend yield of 3-7%: 

1. Merck - PE of 13 with yield of 3.5%
2. Pfizer - PE of 14 with yield of 6.9%
3. Sanofi - PE of 17 with yield of 4.5%
4. Novartis - PE of 17 with yield of 3.1%

 

Thesis is AI makes input costs go down for Big Pharma, but drug prices stay the same or go up. 

 

Over a 10 yr time frame for basket approach: 

• downside is 1.5-2x ? 
• upside is 4x+ ?

Thoughts from group? : 

• on upside or downside guesstimates?
• agreement/disagreement re favourable risk/reward ratio if guesstimates are accurate? 

Overall, many businesses will take advantage of AI. I'm thinking Big Pharma is one where AI is economically feasible. 

 

[winjitsu]

https://blog.google/products/search/gemini-3-search-ai-mode/

 

Gemini 3 rolling out (though at the moment only to the $250 pro subscribers and those with API keys).

 

Benchmarks look promising though again they are very selective by not comparing to Claude Opus and GPT 5.1 Thinking. 

 

Google's cash flow and custom TPUs could be the biggest difference that win them this race. Compared with Chat GPT 5.0 and 5.1 which have been throttling and trying to save on inference when possible. 

[Spekulatius]

19 hours ago, investmd said:

Thesis: AI can be a catalyst for Large Pharma. 

 

AI for drug discovery has long been an obvious application for AI in healthcare. However, instead of early stage biotech being the big winners, it might be the boring old steady Big Pharma like MRK and SNY that are currently trading at P/E ratios of <15 and have steady 4-5% dividend yields. 

 

AI has the potential to drastically decrease cost of both drug discovery & development (clinical trial cost) thus driving down spend at Big Pharma. Then these massive, stable conglomerates can see a significant multiples expansion.  Seems to be upside with downside protection? 

How is AI reducing clinical trial costs?

 

On drug discovery, I think the big story for big pharmaceutical will be licensing compounds from Chinese researchers for way less than they can get similar quality leads from biotechs.  China has quietly become a biotech powerhouse.

 

Merck and Novartis are actively pursuing this path to beef up their pipelines.

Edited November 19, 2025 by Spekulatius

[investmd]

4 hours ago, Spekulatius said:

How is AI reducing clinical trial costs?

 

On drug discovery, I think the big story for big pharmaceutical will be licensing compounds from Chinese researchers for way less than they can get similar quality leads from biotechs.  China has quietly become a biotech powerhouse.

 

Merck and Novartis are actively pursuing this path to beef up their pipelines.

AI can and I believe is reducing cost of clinical trials by patient screening tools to help pick appropriate patients, synthetic control arms, digital twins, optimizing efficiency of CROs (contract research organizations)...

 

and yes - going to China is a way for Pharma to further decrease cost of clinical trials. 

[formthirteen]

6 hours ago, Spekulatius said:

How is AI reducing clinical trial costs?

 

I don't know if NovoScribe is still used by Novo. They could have scrapped it after learning more about EU regulations, limits of AI, and RAGs (recall, non-deterministic, precision, etc):

https://www.claude.com/customers/novo-nordisk

 

Quote

 

Initially the team focused on perfecting NovoScribe for CSR production, and then expanded to device protocol documentation and patient materials.

The time saved by NovoScribe was immediate and substantial. “Claude has helped us cut writing times on CSRs by 90% so we can get documentation directly into human hands for review and approval,” said Waheed Jowiya, Digitalization Strategy Director at Novo Nordisk.

Review-to-approval cycles for these automated CSRs have also decreased significantly as reviewers became increasingly satisfied with clinical interpretation and overall output quality. For new trials, the team can now produce a complete study booklet and patient guide in under a minute — materials that previously required months of back-and-forth work with external agencies.

 

 

Quote

The pinnacle of this process is the clinical study report (CSR), which summarizes drug trial results in a document that can be up to 300 pages long. Producing a CSR was a grueling multi-month cycle of writing, reviews, rewrites, and approvals. Staff writers averaged only 2.3 CSRs per year, and the manual process remained prone to errors and subjective interpretation.

 

[Spekulatius]

5 hours ago, investmd said:

AI can and I believe is reducing cost of clinical trials by patient screening tools to help pick appropriate patients, synthetic control arms, digital twins, optimizing efficiency of CROs (contract research organizations)...

 

and yes - going to China is a way for Pharma to further decrease cost of clinical trials. 

Thx. I do wonder with the above if this is really AI or just having a good digital infrastructure.

 

Regarding China -  China is not just used for clinical trials, the more important step is that Pharma licenses clinical compounds in China at favorable conditions from researchers and then pushes them through clinical trials.  Apparently the leads/ compounds from there are of high quality according to Merck (from CC transcripts).

[rogermunibond]

@beerbaron so stuff like optical networking Lumentum and Innolight?  

[beerbaron]

Yeah I was thinking about the lasers, optical transducers, etc... I don't think those have reached the performance ceiling equivalent to the silicon limit. I'm a bit wary of investing in anything related to AI hype right now but if your are going to scale super-horizontally you need ridiculous latency and bandwidth to avoid bottlenecks. We should start another thread discussing opportunities for picks ans shovels for AI where valuations have not hit stratosphere yet.

 

BeerBaron

[investmd]

17 hours ago, Spekulatius said:

Thx. I do wonder with the above if this is really AI or just having a good digital infrastructure.

 

Regarding China -  China is not just used for clinical trials, the more important step is that Pharma licenses clinical compounds in China at favorable conditions from researchers and then pushes them through clinical trials.  Apparently the leads/ compounds from there are of high quality according to Merck (from CC transcripts).

I think we are in agreement that there are several reasons why cost and time to market for new drugs may come down, thus, increasing earnings for Pharma. Seems like a good time to get into Merck, Pfizer, Sanofi, Novartis...

[Dalal.Holdings]

I just checked and NVDA is all the way back down to where it was in …August. And yet people are talking about the sky falling.
 

It’s incredible how emotionally levered some people seem to be, especially those who partake in these frothy stocks. I guess it’s implicit acknowledgement that there’s little downside protection with these things once they turn—hence the panic

[investmd]

On 11/18/2025 at 11:46 PM, formthirteen said:

 

I don't know if NovoScribe is still used by Novo. They could have scrapped it after learning more about EU regulations, limits of AI, and RAGs (recall, non-deterministic, precision, etc):

https://www.claude.com/customers/novo-nordisk

 

 

 

+1 

 

There seem to be many ways for AI to reduce costs for Big Pharma. Let's see if it eventually bears out in earnings reports

 

Let's move this discussion of AI being a catalyst for Big Pharma to the Investment Ideas section. I'd take a basket approach for the strategy with 4-5 pharma stocks, but will start discussion in a thread for Merck, MRK. 

[MungerWunger]

Ben Evans, AI eats the world slide deck: https://www.ben-evans.com/presentations

[Eldad]

Recently GOOGL goes up 

 

MSFT, ORCL, Softbak, NVDA go down. 
 

Is the market saying Gemini 3 just made Open AI ipo dead on arrival? 

[LC]

3 hours ago, Eldad said:

Is the market saying Gemini 3 just made Open AI ipo dead on arrival? 

Yes and the metric to track is enterprise adoption. My understanding is the Google/ Gemini ecosystem is the most enterprise-ready of the AI LLMs. Actually, Microsoft/OpenAI screwed the pooch by letting Google get a foothold where they had first mover advantage. They just didn't deliver a seamless, enterprise-ready package. 

[NnnnotSoSmart]

AI Roundtable: What Everyone Missed About Gemini 3 

Chapters:
00:00 - The Rise of Gemini 3 and AI Super Intelligence
05:46 - Gemini 3: A Game Changer in AI Technology
08:27 - Benchmarking AI Performance: The Vending Machine Experiment
11:16 - The Future of AI-Driven Businesses
13:58 - Creating with AI: The New Era of Game Development
16:43 - Advancements in Voice Interaction and Translation
19:29 - Google's Competitive Edge in AI Technology
26:38 - AI Shopping Revolution
36:57 - The Future of AI and Biosecurity
43:42 - AI in Bio-Weapon Prevention
53:16 - The Arms Race of AI Safety
56:18 - Understanding Cursor: The AI Coding Assistant
58:50 - Project Prometheus: AI in Engineering and Manufacturing
01:02:37 - The Future of AI: From Superintelligence to Industrial Agents
01:07:06 - Concrete Milestones for an Abundant Future
01:14:36 - The Path to Universal Happiness with AI

 

 

[Spekulatius]

High 5’s for Google everywhere means it’s likely time to take some profits. I still think the risk to Google main search business are there and this gets since it has the dominant market position, the outcome of any change in business is a net negative.

 

Now they can be retro new revenue sources but Search is such a beast that if Google’s market share drops, it’s for going to hit the bottom line.

[DegenerateGambler]

9 hours ago, NnnnotSoSmart said:

AI Roundtable: What Everyone Missed About Gemini 3 

Chapters:
00:00 - The Rise of Gemini 3 and AI Super Intelligence
05:46 - Gemini 3: A Game Changer in AI Technology
08:27 - Benchmarking AI Performance: The Vending Machine Experiment
11:16 - The Future of AI-Driven Businesses
13:58 - Creating with AI: The New Era of Game Development
16:43 - Advancements in Voice Interaction and Translation
19:29 - Google's Competitive Edge in AI Technology
26:38 - AI Shopping Revolution
36:57 - The Future of AI and Biosecurity
43:42 - AI in Bio-Weapon Prevention
53:16 - The Arms Race of AI Safety
56:18 - Understanding Cursor: The AI Coding Assistant
58:50 - Project Prometheus: AI in Engineering and Manufacturing
01:02:37 - The Future of AI: From Superintelligence to Industrial Agents
01:07:06 - Concrete Milestones for an Abundant Future
01:14:36 - The Path to Universal Happiness with AI

 

 

The path to universal happiness with AI is such a bold claim i think it's people annoiting themselves Gods.

[NnnnotSoSmart]

On 11/22/2025 at 1:32 PM, DegenerateGambler said:

The path to universal happiness with AI is such a bold claim i think it's people annoiting themselves Gods.

Agreed.  Artificial Intelligence has the potential to become a modern day Tower of Babel.

[NnnnotSoSmart]

The Structural Collapse: How Google’s Integrated Stack Is Dismantling the OpenAI Thesis
 

A leaked internal memo reveals the tectonic shift reshaping artificial intelligence, where platform economics are defeating venture-backed innovation at the exact moment markets assumed the opposite.

 

https://substack.com/inbox/post/179701197?r=6p7b5o&utm_medium=ios&utm_source=post-publish&triedRedirect=true

 

[NnnnotSoSmart]

On 11/22/2025 at 1:32 PM, DegenerateGambler said:

The path to universal happiness with AI is such a bold claim i think it's people annoiting themselves Gods.

This guy has what I think is a more realistic take:

 

Opinion

 

AI extremists are peddling science fiction

AI realism embraces humility and a basic truth: Technologies succeed when they improve lives.

 

September 12, 2025

 

"This fear is evident whether you are attending a think tank discussion on the future of AI or having a private dinner conversation. The concern heard often in Washington, and sometimes in technology circles, is that if action isn’t taken swiftly, AI might become sentient, break free from human oversight and pose a threat to us — either physically or economically.

 

Right now the AI debate is dominated by two extremes. Doomers believe AI will become godlike and destroy us. Zealots believe AI will become godlike and save us. Their conclusions are different, but their logic is the same: AI will surpass human intelligence and slip beyond human control. Both positions are rooted in science fiction, not science. And both ignore how innovation actually works.

 

There’s a third way: AI realism.

 

AI realism begins with a basic truth: Technologies succeed when they improve lives. The printing press, antibiotics, cars and the internet were not adopted because they were powerful or frightening. They were adopted because they solved problems. AI will follow that same pattern."

...

"Talk of AGI assumes a finish line, a moment when AI suddenly leaps from a tool to a god. But intelligence isn’t a threshold. It’s a spectrum, a constant evolution of abilities. Human knowledge expands every year, reshaping what we call “intelligence.” What seems advanced today becomes irrelevant tomorrow, as each generation of AI models demonstrates. Intelligence is progression without end, not a destination. And that endpoint is unknowable — unmeasurable because no human, no machine, will ever grasp how much knowledge there truly is in a constantly expanding universe. In other words, the meaning of intelligence changes as our knowledge expands. That means AGI is impossible to measure and, therefore, impossible to attain."

 

https://wapo.st/4oYNYGA

 

 

 

Edited November 24, 2025 by NnnnotSoSmart

[investmd]

On 11/18/2025 at 5:36 PM, Spekulatius said:

How is AI reducing clinical trial costs?

 

On drug discovery, I think the big story for big pharmaceutical will be licensing compounds from Chinese researchers for way less than they can get similar quality leads from biotechs.  China has quietly become a biotech powerhouse.

 

Merck and Novartis are actively pursuing this path to beef up their pipelines.

@Spekulatius: Some example to answer q of how AI could be reducing costs of drug development - both discovery and clinical trials: 

Here are 3 private companies with $1B+ valuations that have a thesis of saving costs by trading labor intensive work for compute and thus make R&D more capital efficient: 

• Xaira - raised a $1B round in 2024 to reimagine drug discovery R&D
• Isomorphic Labs - use protein folding knowledge for drug discovery
• Formation Bio - recently raised $370M to make clinical trials more efficient using automation - next generation of CROs

Again my thesis is that multiple companies are using AI to improve time and capital efficiency of drug discovery - all of which benefits Big Pharma. Thus, by decreasing costs earnings at Big Pharma can increase even if revenue stays stable. 

 

Recall that most use the figure of $2B as cost of bringing a new drug to market. Use compute to make that process more efficient and reasonable to expect profits to increase.  

 

Based on the thesis of AI lowering cost of drug discovery, I'm buying a basket of Big Pharma stocks rather than picking one. Relatively safe bets as PEs are approx 15 and they pay out a dividend of 3-6%, but upside exits due to use of AI to decrease costs. 

[Spekulatius]

The AI used here seem more like the old fashioned machine learning (Xaira)  than LLM. LLM’s may be useful for drug submission but the hallucinations intrinsic in LLM may be a huge obstacle here.

I think Incyte about 25 years ago also claimed the same thing about drug discovery but maybe this time the time is right. $1B in funding for a preclinical company is a lot of money.

[WayWardCloud]

If an AI lab is able to massively improve drug discovery, I would expect them to divert a big chunk of the profits away from legacy pharma, turning them into a lower margin OEM, no?

[rogermunibond]

@WayWardCloud the legacy pharma are development, production, sales and marketing powerhouses.  small early stage companies that do drug discovery don't have the muscle with sales and marketing to push new drugs to clinicians and CMS and other nation's health systems.  Most of the legacy pharma have drastically cut down their early stage research pipelines from where they were in the 2000s iirc.  they go shopping for promising startups and early stage companies.  I think they still earn a pretty sizeable return for their distribution prowess.

 

in fact if MAHA has it's way and DTC drug advertising is killed off, the legacy pharmas will have even better returns and bigger moats.