Hey HN, this is David from Aluna (S24). We work with diagnostic labs to build datasets and evals for oncology tasks.

I wanted to share a simple RL environment I built that gave frontier LLMs a set of tools that lets it zoom and pan across a digitized pathology slide to find the relevant regions to make a diagnosis. Here are some videos of the LLM performing diagnosis on a few slides:

(https://www.youtube.com/watch?v=k7ixTWswT5c): traces of an LLM choosing different regions to view before making a diagnosis on a case of small-cell carcinoma of the lung

(https://youtube.com/watch?v=0cMbqLnKkGU): traces of an LLM choosing different regions to view before making a diagnosis on a case of benign fibroadenoma of the breast

Why I built this:

Pathology slides are the backbone of modern cancer diagnosis. Tissue from a biopsy is sliced, stained, and mounted on glass for a pathologist to examine abnormalities.

Today, many of these slides are digitized into whole-slide images (WSIs)in TIF or SVS format and are several gigabytes in size.

While there exists several pathology-focused AI models, I was curious to test whether frontier LLMs can perform well on pathology-based tasks. The main challenge is that WSIs are too large to fit into an LLM’s context window. The standard workaround, splitting them into thousands of smaller tiles, is inefficient for large frontier LLMs.

Inspired by how pathologists zoom and pan under a microscope, I built a set of tools that let LLMs control magnification and coordinates, viewing small regions at a time and deciding where to look next.

This ended up resulting in some interesting behaviors, and actually seemed to yield pretty good results with prompt engineering:

- GPT 5: explored up to ~30 regions before deciding (concurred with an expert pathologist on 4 out of 6 cancer subtyping tasks and 3 out of 5 IHC scoring tasks)

- Claude 4.5: Typically used 10–15 views but similar accuracy as GPT-5 (concurred with the pathologist on 3 out of 6 cancer subtyping tasks and 4 out of 5 IHC scoring tasks)

- Smaller models (GPT 4o, Claude 3.5 Haiku): examined ~8 frames and were less accurate overall (1 out of 6 cancer subtytping tasks and 1 out of 5 IHC scoring tasks)

Obviously, this was a small sample set, so we are working on creating a larger benchmark suite with more cases and types of tasks, but I thought this was cool that it even worked so I wanted to share with HN!

Project Euler

Prebake: A Straightforward Developer Platform for Kubernetes

Let the Mind-Control Games Begin

Vortex – An extensible, state of the art columnar file format

GoDaddy is auctioning a 15-year-old .org from an FOSS volunteer group – help?

New Glenn Mission NG-2

Renowned exoplanet researcher is bringing quest to find explanets back Canada

Astringent flavanol fires locus-noradrenergic system, regulates autonomic nerves

The Duchess Who Invented Science Fiction

NLnet's €21.6M fund for open-source internet projects

Veilid: Distributed Decentralized Framework (From CultoftheDeadCow)

We analyzed 47,000 ChatGPT conversations. Here's what people use it for

Python Strftime Cheatsheet

Convex raises $24M to reinvent back ends

Marble by World Labs: Multimodal world model to create and edit 3D worlds

10× Faster Log Processing at Scale: Beating Logstash Bottlenecks with Timeplus

Northern Lights Dazzle U.S. Skies After Powerful Solar Storm

Teradar raises $150M for a sensor it says beats lidar and radar

Planchón-Peteroa volcano enters new eruptive phase

Arch-delta Saves 80% Of Bandwidth On Upgrades

Haiku Activity and Contract Report, October 2025

The Al Bubble Is Worse Than You Think [video]

Project OSSAS: Custom LLMs to Process 100M Research Papers

LAION Dataset Explorer

Google launches a lawsuit targeting text message scammers

Helm v4.0.0

Show HN: Cancer diagnosis makes for an interesting RL environment for LLMs

Show HN: AI music discovery for super nerds (Now on iOS AND macOS)

Extracting Playable Instrument Models from Short Audio Examples

DSearch: .NET IQueryable Dynamic Search Library

Show HN: SpendScope – Track AI Costs Across OpenAI, Anthropic, Google AI