No‑code tools shift skills

Drug research is getting easier to run without code, but the hard part is shifting to deciding which biology questions are worth testing. (reuters.com) (computerweekly.com) On April 14, Amazon Web Services launched Amazon Bio Discovery, a tool for early-stage drug research that lets scientists run complex computational workflows without writing code. Reuters reported the system is aimed at speeding discovery work by making those workflows easier to use. (reuters.com) (aboutamazon.com) Drug discovery starts with a simple problem: researchers must guess which molecules might affect a disease, then test those guesses in the lab. Amazon said its platform gives scientists access to more than 40 biology models, helps choose settings, and sends shortlisted candidates to contract lab partners for synthesis and testing. (aboutamazon.com) (aws.amazon.com) Novo Nordisk announced its own OpenAI partnership on April 14, saying it will use the technology across drug discovery, manufacturing and commercial work. The company said pilot programs start immediately in research and development, manufacturing and commercial teams, with broader integration planned by the end of 2026. (cnbc.com) (pharmaceutical-technology.com) Those announcements point to the same purchase decision: companies are paying for easier handoffs between software teams, lab teams and decision-makers. Amazon’s own description says the product is meant to connect “dry lab” computing work with “wet lab” experiments in one application, reducing delays from manual handoffs. (aws.amazon.com) (aboutamazon.com) That changes which skills carry the most weight. If software can generate candidates, tune models and package results, researchers still need to judge whether a target makes biological sense, whether an experiment is well designed, and whether the next team can act on the output. (aboutamazon.com) (reuters.com) Big drugmakers have been adding artificial intelligence tools for years, but the pitch has widened from molecule design to the whole workflow. Reuters said Novo wants OpenAI’s tools to analyze complex datasets, identify promising drug candidates and improve efficiency in manufacturing, supply chains, distribution and corporate operations. (reuters.com) (cnbc.com) Amazon is making a similar argument from the software side. Its launch materials say scientists can use built-in agents to select models, optimize configurations and evaluate candidates, then feed lab results back into the next design cycle. (aboutamazon.com) (aws.amazon.com) Neither announcement says the software replaces biologists. Both describe systems that narrow options and speed iteration, while people still choose the disease target, define success, interpret noisy results and decide which failures are worth learning from. (reuters.com 1) (reuters.com 2) As more of the technical setup disappears behind buttons and prompts, the bottleneck moves to judgment: asking better questions, designing cleaner experiments and explaining results across teams that still have to turn a model output into a drug program. (aws.amazon.com) (aboutamazon.com)