Winton Gibbons

@VPrasadMDMPH Science always seems most likely to fall prey to “those who don’t learned history are doomed to repeat it.

This is a cancer therapeutic vaccine. Over 40 years we have trialed dozens of them. All* failed. This is in early phase trials. When tested in a large randomized trial, It will also fail. *Provenge deprived control arm pts of docetaxel ftr

@VPrasadMDMPH @Amgen I heard something similar at an on site AbbVie conference. The company showed visible anger to the academic scientists assembled.

When @Amgen tried to replicate 53 papers, they could only do this 11% of the time Stand up for science would mean addressing this But everyone is incentivized not to care

@DTWillingham Agreed. The same prompt (if it’s for a real task) will return different results, often quite different. And certainly OpenAI is always tinkering, obviating even reproducible approaches.

Amidst all the AI rhetoric, the talking point I'm most sick of: "If you didn't get a useful answer from Chat-GPT, it's because you wrote a bad prompt." I don't even care if it's true, just shut up.

@KonstantinKisin Indeed

For my British and European friends who are "shocked" and "surprised", here are 10 reasons you didn't see this coming. Read this short post and then read the replies from our American friends who will confirm what I'm saying. 1. Americans love their country and want it to be the best in the world. America is a nation of people who conquered a continent. They love strength. They love winning. Any leader who appeals to that has an automatic advantage. 2. Unlike Europeans, Americans have not accepted managed decline. They don't have Net Zero here, they believe in producing their own energy and making it as cheap as possible because they know that their prosperity depends on it. 3. Prices for most basic goods in the US have increased rapidly and are sky high. What the official statistics say about inflation and the reality of people's lives are not the same. 4. Unlike you, Americans do not believe in socialism. They believe in meritocracy. They don't care about the super rich being super rich because they know that they live in a country where being super rich is available to anyone with the talent and drive to make it. They don't resent success, they celebrate it. 5. Americans are the most pro-immigration people in the world. Read that again. Seriously, read it again. Americans love an immigrant success story. They want more talented immigrants to come to America. But they refuse to accept people coming illegally. They believe in having a border. 6. Americans are sensitive about racial issues and their country's imperfect history. They believe that those who are disadvantaged by the circumstances of their birth should be given the opportunity to succeed. What they reject, however, is the idea that in order to address the errors of the past new errors must be made. DEI is racist. They know it and they reject it precisely because they are not racist. 7. Americans are the most philosemitic nation on earth. October 7 and the pro-Hamas left's reaction shocked them to their very core because, among other things, they remember what 9/11 was like and they know jihad when they see it. 8. Americans are extremely practical people. They care about what works, not what sounds good. In Europe, we produce great writers and intellectuals. In America they produce (and attract) great engineers, businessmen and investors. Because of this, they care less about Trump's rhetoric than you do and more about his policies than you do. 9. Americans are deeply optimistic people. They hate negativity. The woke view of American history as a series of evils for which they must eternally apologise is utterly abhorrent to them. They believe in moving forward together, not endlessly obsessing about the past. 10. America is a country whose founding story is one of resistance to government overreach. They loathe unnecessary restrictions, regulations and control. They understand that freedom comes with the price of self-reliance and they pay it gladly.

PinkDx sounds like it is getting it right from the start: #clinicalrigor #multiomics #machinelearning Also home their is an early focus on #marketaccess #reimbursement Good luck!! prnewswire.com/news-releases/…

Payors, Labs Parsing State Biomarker Testing Laws to Better Understand Coverage Mandates. Laws passed in 14 states appear to require commercial payors to cover biomarker tests broadly that meet certain evidentiary criteria, but payors may have a... ow.ly/lqO0104YLcz

Rather #counterintuitive “Chat” on #AGI #superAGI chat.openai.com/share/3e78a71f… #STEM #biotechnology #lifesciencetools #cancer #medicine

@AppleHelix Agreed. How on earth would those studies be funded? Is there some "if, then;" whereby, the USPTO says if a study meets some particular endpoint that the patent is then automatically granted?

Also, I am all for novel ways to improve Pharma R&D productivity. I think Big Pharma should be more open-minded. But how would your recommendation #1 even work? Does this mean you want USPTO to hold off awarding a patent until we get 2 positive Phase 3 readouts?

I am throwing down the gauntlet. Happy to debate all of the authors on this paper live on Twitter Spaces. @Aris_Angelis @ojwouters @ElsTorreele @martinmckee Are you guys up for it? @PeterKolchinsky @daphnezohar @sciencescanner you guys want to join me?

@JAMA_current Shape matters, not just AUC, in many cases more. Why are there no ROC curves shown? Also, the use of ROC curves, which are prevalence independent, ignores positive ("rule-in") and negative ("rule-out") predicitve values. Without those, AUCs, low or high, cannot really be judged.

In this retrospective study of predictive accuracy, existing stroke-specific risk prediction models and novel machine learning techniques did not significantly improve discriminative accuracy for new-onset stroke compared with the pooled cohort equations. ja.ma/3j6zDef

#ChatGPT fatally flawed by propagating (expands?) scientific publication bias (#GPT-3 itself agrees a huge problem). Need severe warning, not just “I might be wrong.” Betting it leads to stronger status quo, less new knowledge. #retraction #publicationbias #reproducibilitycrisis

plos.org/reproducibilit… Useful for #IVD R&D. Many studies replicate poorly, and this delays progress, both scientifically and economically. #precisionmedicine #AACC #Dx #omics #cancer #biomarkers #companiondiagnostics #CDx #proteomics #bioinformatics #genomics #medtech #biotech

European Investment Bank Finances Innovative Diagnostics Developer #Numares Health with €20 Million Numares develops precision multi-biomarker diagnostic tests for patients with chronic kidney, heart and liver disease. Numares.com bit.ly/3FKCD7s

#Numares and #MayoClinic deepen its collaboration for patient care in #metabolic & #inflammatory conditions as #Mayo invests in numares with a convertible equity investment. Together, we work to create #NMR #metabolomic tests to help physicians better care for their patients.

numares is committed to #kidneycare nationally and locally near our US headquarters in #TheWoodlands, Texas.

What a great addition to the #numares Supervisory Board. Greg brings decades of #diagnostics and #omics experience from #Illumina and #Roche.

numares continues to drive #metabolomics and #multiomics into routine clinical use. Easy-to-use NMR. Great employees enabling.

#numaresNews: Detection of #MS disease progression is still a diagnostic gap. Excellent scientific work of @UniofOxford and #numares #NMR #diagnostics combining #biomarkers may change how #multiplesclerosis #patients are diagnosed & treated: Read more: lnkd.in/eApvgsxJ

@AppleHelix The data is so dramatic and "perfect" that perhaps the hypothesis should more likely be an issue with sample or data handling.

That is a pretty sharp drop off after 80. It's almost as if a biological switch was flipped. We need to look for the 80+ outliers who maintain high titers and understand what's different about them. This is an anti-aging play.

Here is another positive milestone for #numares as we continue to build momentum. We also expect to file another 510k application to the #FDA this year for our #kidney function assay. #metabolomics #precisionmedicine #invitrodiagnostics #cardiacdisease numares.us/en/company/med…