🪴 Oliver’s Note - AI for biomedical research
Did you see the announcement from Anthropic that Claude is now available for Life Sciences?
With Claude Sonnet 4.5 new workflows based on enhanced connectors to Benchling, BioRender, PubMed, Wiley Scholar Gateway, Synapse, and 10x Genomics, are possible.
New agent skills enable task-specific, protocol-driven performance, starting with single-cell RNA-seq QC, while scientists can build custom Claude skills for repeatable workflows.
The model supports literature reviews, hypothesis generation, protocol generation, bioinformatics and data analysis, and clinical/regulatory documentation, backed by a growing library of prompts and hands-on support from Anthropic’s Applied AI teams and cloud partnerships; the AI for Science program offers free API credits to researchers and invites project submissions to accelerate life-sciences discovery from research to commercialization.
And this is only the beginning - check out the roadmap on GitHub. A staggering number of skills and connections are coming!
This looks like a very helpful, broadly based toolkit that will allow scientists to build AI into their existing workflows, accelerating progress by helping with tasks that LLMs are very good at, such as textual analysis and managing vast amounts of information.
But, not so fast…
LLMs are only one kind of intelligence and may well not be the path to Artificial General Intelligence that has been touted by some of the big companies. I learned this from Dr. Gary Marcus who was part of a panel this past weekend at my daughter's university, Hampshire College in Massachusetts. The sessions focused on the future of higher education so AI was only a part of what as discussed, but it was the first time that I had heard the limitations of LLM chatbots articulated to clearly. According to Marcus LLMs represent only one mode of intelligence that humans are capable of. In his view, the following kinds of intelligence, required for human-like AGI, are missing:
True Reasoning and Logic: Marcus argues that LLMs are masters of statistical approximation but cannot perform genuine logical or causal reasoning. They can't reliably figure out why something happens or deduce solutions to novel problems that don't resemble their training data. He often uses their failure at simple logic puzzles (like "river crossing" problems) as an example.
World Models and Common Sense: He stresses that humans operate with a robust, built-in "common sense" understanding of the world—we know how objects interact, what people's motivations are, and the basic rules of physics. He argues LLMs lack this internal model of the world and are just "stitching variations on text that it has seen." He used chess as an example, claiming that LLMs can faithfully explain the rules of the game, but when actually playing will disregard the rules, demonstrating that they have not actual knowledge of the game or a model of it in mind.
Symbol Manipulation: Marcus is a major advocate for "neurosymbolic" or "hybrid" AI. He believes true intelligence requires combining neural networks (the pattern matchers) with classical, symbolic AI—systems that can manipulate abstract concepts, variables, and rules, much like algebra or a computer program.
Abstraction and True Innovation: He argues that what LLMs produce isn't true innovation but sophisticated imitation. A human can invent a new game, like chess, from scratch, complete with its own novel rules and abstract goals. An LLM, in his view, can only generate new content (like a chess game or a story) based on the patterns of the many games and stories it has already ingested.
So, while you try out Claude for Life Sciences, rest assured that there is still much work to be done before we have AGI.
🗓️ Upcoming Grants
Standard NIH Due Dates
For new R01 submissions these are Feb 5, June 5, Oct 5
For R01 renewal, resubmission or revisions these are Mar 5, July 5, Nov 5
For more information check out the NIH Standard Due Dates page.
Damon Runyon Scholars Program for Advancing Research and Knowledge
The Damon Runyon SPARK program offers a one-year post-baccalaureate internship for current seniors and recent graduates, including those from underrepresented backgrounds, to develop research skills in cancer biology, with the goal of progressing to graduate or medical school. Scholars are placed in host laboratories, receive mentoring and programmatic activities (retreats and virtual sessions), and obtain a total award of up to $61,000 to the host institution plus a small community-building allowance for program directors; reporting, ethics, and compliance requirements accompany the funding.
Notice link
eligibility: current college seniors and those who have graduated with a bachelor’s degree within the last 36 months - more criteria at link
deadline: February 2nd, 2026
funding amount: up to $61,000
Damon Runyon Cancer Research Foundation
🦋 Growth Mindset: Quantifying Attitudinal Energy and Consciousness
We all experience energy fluctuations, feeling either peaceful and connected, or exhausted and frustrated, especially when unfocused. What if we had a framework to understand these energy shifts and a tool to evaluate our individual energy profile?
I’ve been learning one approach to this issue: iPEC’s Energy Leadership Index® (ELI). This is an attitudinal assessment that quantifies an individual’s energy mix, both in the absence or presence of stress. The schema categorizes energy and how we think, feel and act into 7 Levels that range from the catabolic to the anabolic. These terms, borrowed from metabolism have specific meanings:
Catabolic energy (Levels 1-2) is draining and destructive, linked to stress and fear. Level 1 is victimhood; Level 2, conflict. Can help you get through tough moments.
Anabolic energy (Levels 3-7) is constructive, growth-oriented, and healing. Higher levels relate to power and satisfaction, shifting from self-focus to reconciliation and non-judgment.
Though initially skeptical as an older scientist, I've found this schema and its test surprisingly helpful. It's not magic, but it aids in self-awareness and facilitates personal growth.
Interested? You know where to find me.
🏆 Success Tools - Grant Webinars
Starting October 30th I am giving four weekly free webinars on grant writing.
I am covering the Specific Aims page, writing the Research Strategy, understanding Review and Reviewers and how to Resubmit.
You can register here for individual webinars or subscribe to the series.
Why not share this information with a colleague or members of your lab?
Book a meeting via my website www.bogler.cc
Send an email to me at: [email protected]