I like your idea of making difficult to read literature more accessible to the public. It seems like a step in the right direction.
As for the problem "It's no longer possible to believe clinical research" the fundamental problem is that clinical research experiments are not repeated. Edison repeated experiments frequently. Science is not based on trust. Almost all published papers are basically just "Trust me bro this is what I did and this is what it means". Anonymous peer review is simply "these smart guys you don't know say the author is being straight up". That is not science. You describe your experiment in detail so others can repeat it and verify the results for themselves. That is science.
Medical experiments cannot be repeated for two reasons. First they are too expensive. They cost tens of millions of dollars. No one has the resources to repeat a drug trial and verify the results. Secondly, even if you were somehow to gather up the tens of millions of dollars, the ethics committee will forbid you from repeating the experiment. Why? They will argue that you are denying the placebo group a drug of known efficacy and that is immoral. So not only are drug experiment not science, they can never be science under the current constraints.
Since we cannot repeat the experiments we are stuck with "Trust me bro, this is what I did and this is what it means". I think we can all see the problem with that now.
My observation is far more dark than this - the real issue is that the majority of our research community are simply all to willing to be intentionally dishonest to serve A) the desired outcome of their funding sponsors, and/or B) to promote their (almost always) hard left ideological political agenda.
Today's clinical research is 'too expensive' to repeat due to the massive amounts of paperwork and the need to run very large and 'controlled' trials in order to reach statistical significance...and perhaps do this twice. Smaller, adaptive, and objectively monitored studies can shed a lot of light on the reality of the safety and effectiveness of these products. And using good post-marketing, or staged-release studies, with objective monitoring and analyses, will provide the kind of data we are mainly missing when a new product is first approved...longer term data, at a variety of dose ranges, in a broader group of individuals. These data (although 'messy' versus the data from the original well-controlled studies) could be collected and analyzed in 'real time' using current technology and analysis methods. With regard to ethics...I think placebo arms would not be necessary in the types of studies I am envisioning.
The ethics argument in regards to placebo arms in such RCTs does not hold water. After all, the intervention, even if showing early benefits, is not and will not be available to the general pop. before some sort of use authorization. Therefore, the placebo arm is just on par with the rest of the population.
IMHO, in the Pfizer trial (and quite a few other trials), this is simply a tactic to make sure that the collected data is cut short.
Worse: Once an intervention has been approved, it truly would be unethical to conduct a study including a placebo arm. But this is a circular type of reasoning.
The whole immoral line is bullshit. I first heard that with the covid vaccines and giving them to the control group, therefore destroying the control group. I have found out since that it is common sense. It's pharma who is immoral pushing this crap in the way it does, not conducting real safety experiments on it's products.
I have been in the pharm and medical device development field for over 50 years. I have gone from the most basic positon of medical writer to director of clinical research, quality assurance and regulatory affairs. I read your article and it appears you missed your own point...that the data and pubs are tainted. That is absolutely true. But your solution does not address that issue. It is good for the 'average' reader to get a good abstract of a publication or series of publications, but that does not solve the problem you identified.
Thanks for your comment. You're absolutely right by the way, but I believe you're only correct in this exact moment.
I had a section about the work that Alexandros Marinos is doing but I took it out. And Chris Martenson, Jessica Rose, JikkyLeaks, and a whole host of ad-hoc peer reviewers. It's an incredible amount of work they put in, and by the time the problems are highlighted, the horse has bolted.
I think that a finely tuned language model like Microsofts will ultimately be able to pick out statistical misdirection at scale. Things like the chicanery over what the placebo was, when it was given, and to whom, and how the control groups get tainted, these will eventually be detectable.
Yes, it will create a game of cat and mouse, but we're already playing that game as we speak. This is just the logical next step.
Secondly, even just a real time dynamic synthesis of 'the data' can have an immediate effect. What do I mean by that? Well take a look at the Ivermectin saga. Whether you believe it works or not (I am strongly of the view that it does), it's hard to argue that to the public the drug was tarnished with the 'there's no evidence that it works' brush.
The horse dewormer/vetinary thing was very effective. What tool stood in the way of that avalanche of misinformation? Ivmmeta was there. An absolutley fantasitc labour of love by some anonymous heroes. Whether you accepted the data or not, the app acted as an impossible stick in the mud for 'the narrative'.
We got lucky that someone, or some people, somewhere, worked tirelessly and manually to put that thing together. In the not too distant future I think its going to be possible to run projects like that automatically, at scale, by semantically extracting data out of science papers.
That, in the least, takes away the 'theres no evidence it works' tool. We'll no doubt see other tactics then adopted, but we can be ready and nimble to deal with them.
Having designed, executed, analyzed, and presented clinical studies to government agencies...I still think there is a lot of information that cannot be scraped from a 'reading' of a publication or preprint. For example, where would one see the inclusion/exclusion criteria and how they are changed throughout the study in some cases?
In the case of the TOGETHER trials, those protocols were actually published. They were registered on an NIH website, and all versions of them could be read and re-read. Take for example the hydroxychloroquine study I wrote about, that dose was included in the protocols of the study. Even now, I think it will be possible for an AI to check that dose and say that it is toxic.
The protocols were also published for the Pfizer trials. Lets say that a paper *is* published without that inclusion/exclusion criteria, or without a detailed protocol, it will be possible to flag that very clearly on the papers. The public can then see, at a glance, that the publishers haven't provided key information.
That would be very helpful. But my position is that we need to go deeper into the origin story or these 'therapeutics'. For example, I was presenting to an FDA director a new oncology protocol and he suggested I add a third arm to the study using an existing 'proven ' drug. I asked why and he said that the original approval was flawed. So...just saying...there are probably a lot of 'less than effective and safe' therapeutics on the market that are supported by 'the data'.
Oh I agree entirely. We need to go very deep indeed. One big component of all this is the sheer volume of data is too much for one human to handle, but an AI can 'hold in memory' absolutely masses of data at any given moment.
I think we'll be surprised by what can 'become known' even with data that's already public.
That could catalze the political and social will to 'go deep' into the real problem areas in this bizarre industry via subpoenas and select committes.
Phil: I just now ran across your dialogue with Richard Hunter, which I appreciate very much because I had the same question he had. Your analogy to William Tyndale only works if science is generally trustworthy, as the Bible is.
Here is what I wrote to my email list of like-minded friends yesterday after reading your piece. The reason I'm sending it to you is because you may want to publish a clarification as to the reasoning and purpose of your work, as you have with Richard:
JList Friends: I completely honor what Phil Harper is trying to achieve here. I guess I’m just skeptical that his solution will work.
The real problem, as I see it, is not that we need more translations of science papers, or more conservative alarm ringing, or for that matter, more JLists, Epoch Times, or Dennis Pragers, though such are (in many cases—I’ll let you decide 😊) critically valuable, if not indispensible.
The real problem is rapidly encroaching and enveloping secularism. (That is what we need the above to keep reminding us about.) And when a people loses God, it loses everything important to a functioning society, as well as to science itself—common sense, decency, honesty, trained and sensitive consciences, work ethics, everything.
No amount of reworking of science articles so third-graders can understand them is going to help a great deal with the problem of untrustworthy science, as I see it.
But good on Phil Harper for trying. We need people on every one of these battlefronts.
JLister and friend Dan [a politically liberal pastor and evangelist], if you’re still reading here: You are no conservative, or you still don’t know it yet, or you haven’t told us you’ve converted. But you are at the forefront of the most important battle of all: bringing people back to God.
my 2 cents-i am not a credentialed pharma person but I know experimental design and the discipline of the scientific method. In my reading it seems pharma especially is guilty of such things as p-hacking by adding/dropping "problematic" subjects, failing to publish findings for all the original hypotheses, playing games with absolute vs. relative effectiveness and risk, redefining side-effects midway thru etc. many of those slights of hand are in the detail records as I understand it but often not published or available to the FDA. I no longer trust this system, from A to Z. Love PHil's efforts however.
That's the problem. Although there is a big 'translation' problem that the author is trying to address. Others are trying to 'open source' the original data and its analysis...just like the FDA tries to do...but the pharma/etc companies will not allow that for 'patent' and 'business' protection reasons. I once joined an effort to set up clinical trials of 'approved' drugs to verify their effectiveness and safety...but that company changed course. That would be a start, if someone could get sufficient funding.
I lack your expertise in this field. But if I understand this article, a basic benefit of AI analysis is the ability to summarize medical studies as either -
#1 - “This data yields widely varying results depending on who funded the study”
VS
#2 - “This data yields mostly consistent results which create a clear course of action”.
Summary #1 creates immediate “public shaming” of compromised studies, and will improve accuracy through peer pressure.
Summary #2 tells doctors and patients they are at least on the right path and not randomly guessing.
I’m excited for ANYTHING that looks like a path forward
well, there's also the wonderful gift this is to college students. this takes cheating to steroidal levels. where was this when I was a student?! /sarc
Bringing a scientific paper down to understandable levels (for your "average" person) is a great goal. My questions are-- will AI be able to objectively tell the reader if the study has any basic inbred flaws in its methodology, are the conclusion drawn in sync with the facts in the study body and are any of the authors influenced (financially or otherwise) by Big Pharma, government agencies, etc. Reading, interpreting, understanding these papers are no small feat. Ask somebody as knowledgeable as Pierre Kory.
This is a great idea. I just hope it will work as advertised. The ability to get these studies out to average people at an understandable level is very important to allowing people to protect and guide their own medical treatments. Thanks for your efforts, Phil.
As someone with one of those pesky autoimmune problems for which no one can identify the culprit, I have been studying “medical studies, journals, etc...” for at least 15 years and the biggest problem I always encounter is that none of us live in a bubble. I completely understand that we need to do studies for the larger population, but until we admit there are just so many other factors many are bringing to the study, nothing will really be helpful. In my case the research and understanding on the ankylosing spondylitis I have had for 40+ years is ever evolving and at the same time quite stagnant. (As an example, so many of the “other symptoms” of AS which happen in less than 20% of those of us with the disease have happened to me while I have no family history, I have never been overweight, never smoked, am healthy otherwise and yet I have experienced these “rare” percentage other issues - everyone of them??? Probability speaks against this) The data input will always be missing very important information that can never really be extracted on a large scale. We must listen to our own bodies as everyone is different and if we pay attention, what we can’t heal through life changes as we attempt to understand the broken pathways, we can try to mitigate. And really, feeling like you are doing something helps mentally which helps physically as well.
while its nice to have a lay translation... how can we know if its reliable in the first place?reading cliff notes on a false narrative doesnt help.. great idea tho
Thank you for creating this tool. It's much needed in this war for truth and for our freedom from those who would lord over conclusions.
I wonder if we could work together. I have built an evidence sharing platform in which summaries like these are organized into a tree of empirical inquiry with a yes/no question at the root. Our first debate was about whether masks mitigate the risk of COVID-19, and contains an amalgamation of evidence on both sides sorted by user-weighed strength, each open to rebuttal. My goal is to create a space where controversial questions can be examined and where all the evidence is visible, understandable, and tested, with no suppression of either side.
I've hired writers to help me summarize evidence and this has been fruitful both for creating new content and for shaping how the inquiries are organized. However, this is slow and expensive, eating into my savings, and probably won't scale. AI like yours can provide an enormous benefit by allowing even my end-users to request summaries of evidence they themselves submit. This puts the power of knowledge and argument back into the hands of society at large and away from the 'speakers of science.'
Would you like me to share a link to this project?
I think AI will play an important role in finding and summarizing evidence. The parsing of the tree is something I'd still like to leave to the individual. My philosophy for Evincer is to give humans easy access to all the best evidence so that they can decide what's true.
Evincer is an open platform and users often just submit links without summarizing for the reader. So I would like an API that Evincer could access that would help users easily flesh out the missing details. It would accept a link and a debate title: (e.g. "Did Human GHG Emissions Cause Global Warming?") and reply with how the paper addresses the question.
What's the roadmap look like for your project? btw I haven't been able to connect lately. Got a 504 today.
I haven't been able to access the openpaper.sciemce either - I'm not sure what Phil is doing with it, nor how it's being hosted. I also am getting the 504 error. I will go check out your Evincer post - but first skim I do think you could partner up your platform with Phil's AI summary tool.
Outstanding idea, Phil. Even having a postgraduate science degree it is difficult for me at times getting through a lot of advanced science papers.
😱 I did ask ChatGPT, in one session with it, to write a 200 word essay on George Washington at a 7th grade level. It complied very well. AI, like so many advances in science and technology will be used for good and evil. I suspect homework assignments of the future will not be as educational for the student as they once were.
I had found all the evidence of corruption you mention in the beginning of your article and more, it's sooo bad. As a retired RN/RM I can understand the gist of some papers but cannot understand the intricacies. Your tool will be invaluable to many. The interest in looking at scientific papers has grown over the last 3 crazy years. I think your tool will be popular.
Well done, Phil, add this to your references? Keep safe and free.
Episode 353 - The Crisis of Science Corbett Report 02/23/2019
In recent years, the public has gradually discovered that there is a crisis in science. But what is the problem? And how bad is it, really? Today on The Corbett Report we shine a spotlight on the series of interrelated crises that are exposing the way institutional science is practiced today, and what it means for an increasingly science-dependent society.
9:30 pm Eastern Standard Time when I tried to register got an error message. I’ll be patient. 😁 Looking forward to it. I live in the medical literature in my “day job.” I’d like to believe that most of my last 20+ years of research has had honest and transparent findings (some of which have been in those 2 big name journals you referenced). Journals don’t like to publish lots of clinical trials that failed. But with requirements to post all study findings within 12 months of completion, you should have more data for AI to learn.
Hi Phil, very excited about your work, I just tried to register for openpaper.science, and it is not letting me register, wondering if this problem will be fixed soon.
I like your idea of making difficult to read literature more accessible to the public. It seems like a step in the right direction.
As for the problem "It's no longer possible to believe clinical research" the fundamental problem is that clinical research experiments are not repeated. Edison repeated experiments frequently. Science is not based on trust. Almost all published papers are basically just "Trust me bro this is what I did and this is what it means". Anonymous peer review is simply "these smart guys you don't know say the author is being straight up". That is not science. You describe your experiment in detail so others can repeat it and verify the results for themselves. That is science.
Medical experiments cannot be repeated for two reasons. First they are too expensive. They cost tens of millions of dollars. No one has the resources to repeat a drug trial and verify the results. Secondly, even if you were somehow to gather up the tens of millions of dollars, the ethics committee will forbid you from repeating the experiment. Why? They will argue that you are denying the placebo group a drug of known efficacy and that is immoral. So not only are drug experiment not science, they can never be science under the current constraints.
Since we cannot repeat the experiments we are stuck with "Trust me bro, this is what I did and this is what it means". I think we can all see the problem with that now.
My observation is far more dark than this - the real issue is that the majority of our research community are simply all to willing to be intentionally dishonest to serve A) the desired outcome of their funding sponsors, and/or B) to promote their (almost always) hard left ideological political agenda.
Today's clinical research is 'too expensive' to repeat due to the massive amounts of paperwork and the need to run very large and 'controlled' trials in order to reach statistical significance...and perhaps do this twice. Smaller, adaptive, and objectively monitored studies can shed a lot of light on the reality of the safety and effectiveness of these products. And using good post-marketing, or staged-release studies, with objective monitoring and analyses, will provide the kind of data we are mainly missing when a new product is first approved...longer term data, at a variety of dose ranges, in a broader group of individuals. These data (although 'messy' versus the data from the original well-controlled studies) could be collected and analyzed in 'real time' using current technology and analysis methods. With regard to ethics...I think placebo arms would not be necessary in the types of studies I am envisioning.
The ethics argument in regards to placebo arms in such RCTs does not hold water. After all, the intervention, even if showing early benefits, is not and will not be available to the general pop. before some sort of use authorization. Therefore, the placebo arm is just on par with the rest of the population.
IMHO, in the Pfizer trial (and quite a few other trials), this is simply a tactic to make sure that the collected data is cut short.
Worse: Once an intervention has been approved, it truly would be unethical to conduct a study including a placebo arm. But this is a circular type of reasoning.
Thanks for the response Richard!
The whole immoral line is bullshit. I first heard that with the covid vaccines and giving them to the control group, therefore destroying the control group. I have found out since that it is common sense. It's pharma who is immoral pushing this crap in the way it does, not conducting real safety experiments on it's products.
I have been in the pharm and medical device development field for over 50 years. I have gone from the most basic positon of medical writer to director of clinical research, quality assurance and regulatory affairs. I read your article and it appears you missed your own point...that the data and pubs are tainted. That is absolutely true. But your solution does not address that issue. It is good for the 'average' reader to get a good abstract of a publication or series of publications, but that does not solve the problem you identified.
Thanks for your comment. You're absolutely right by the way, but I believe you're only correct in this exact moment.
I had a section about the work that Alexandros Marinos is doing but I took it out. And Chris Martenson, Jessica Rose, JikkyLeaks, and a whole host of ad-hoc peer reviewers. It's an incredible amount of work they put in, and by the time the problems are highlighted, the horse has bolted.
I think that a finely tuned language model like Microsofts will ultimately be able to pick out statistical misdirection at scale. Things like the chicanery over what the placebo was, when it was given, and to whom, and how the control groups get tainted, these will eventually be detectable.
Yes, it will create a game of cat and mouse, but we're already playing that game as we speak. This is just the logical next step.
Secondly, even just a real time dynamic synthesis of 'the data' can have an immediate effect. What do I mean by that? Well take a look at the Ivermectin saga. Whether you believe it works or not (I am strongly of the view that it does), it's hard to argue that to the public the drug was tarnished with the 'there's no evidence that it works' brush.
The horse dewormer/vetinary thing was very effective. What tool stood in the way of that avalanche of misinformation? Ivmmeta was there. An absolutley fantasitc labour of love by some anonymous heroes. Whether you accepted the data or not, the app acted as an impossible stick in the mud for 'the narrative'.
We got lucky that someone, or some people, somewhere, worked tirelessly and manually to put that thing together. In the not too distant future I think its going to be possible to run projects like that automatically, at scale, by semantically extracting data out of science papers.
That, in the least, takes away the 'theres no evidence it works' tool. We'll no doubt see other tactics then adopted, but we can be ready and nimble to deal with them.
Having designed, executed, analyzed, and presented clinical studies to government agencies...I still think there is a lot of information that cannot be scraped from a 'reading' of a publication or preprint. For example, where would one see the inclusion/exclusion criteria and how they are changed throughout the study in some cases?
In the case of the TOGETHER trials, those protocols were actually published. They were registered on an NIH website, and all versions of them could be read and re-read. Take for example the hydroxychloroquine study I wrote about, that dose was included in the protocols of the study. Even now, I think it will be possible for an AI to check that dose and say that it is toxic.
The protocols were also published for the Pfizer trials. Lets say that a paper *is* published without that inclusion/exclusion criteria, or without a detailed protocol, it will be possible to flag that very clearly on the papers. The public can then see, at a glance, that the publishers haven't provided key information.
That would be very helpful. But my position is that we need to go deeper into the origin story or these 'therapeutics'. For example, I was presenting to an FDA director a new oncology protocol and he suggested I add a third arm to the study using an existing 'proven ' drug. I asked why and he said that the original approval was flawed. So...just saying...there are probably a lot of 'less than effective and safe' therapeutics on the market that are supported by 'the data'.
Oh I agree entirely. We need to go very deep indeed. One big component of all this is the sheer volume of data is too much for one human to handle, but an AI can 'hold in memory' absolutely masses of data at any given moment.
I think we'll be surprised by what can 'become known' even with data that's already public.
That could catalze the political and social will to 'go deep' into the real problem areas in this bizarre industry via subpoenas and select committes.
Phil: I just now ran across your dialogue with Richard Hunter, which I appreciate very much because I had the same question he had. Your analogy to William Tyndale only works if science is generally trustworthy, as the Bible is.
Here is what I wrote to my email list of like-minded friends yesterday after reading your piece. The reason I'm sending it to you is because you may want to publish a clarification as to the reasoning and purpose of your work, as you have with Richard:
JList Friends: I completely honor what Phil Harper is trying to achieve here. I guess I’m just skeptical that his solution will work.
The real problem, as I see it, is not that we need more translations of science papers, or more conservative alarm ringing, or for that matter, more JLists, Epoch Times, or Dennis Pragers, though such are (in many cases—I’ll let you decide 😊) critically valuable, if not indispensible.
The real problem is rapidly encroaching and enveloping secularism. (That is what we need the above to keep reminding us about.) And when a people loses God, it loses everything important to a functioning society, as well as to science itself—common sense, decency, honesty, trained and sensitive consciences, work ethics, everything.
No amount of reworking of science articles so third-graders can understand them is going to help a great deal with the problem of untrustworthy science, as I see it.
But good on Phil Harper for trying. We need people on every one of these battlefronts.
JLister and friend Dan [a politically liberal pastor and evangelist], if you’re still reading here: You are no conservative, or you still don’t know it yet, or you haven’t told us you’ve converted. But you are at the forefront of the most important battle of all: bringing people back to God.
Good on you, too.
Janine
my 2 cents-i am not a credentialed pharma person but I know experimental design and the discipline of the scientific method. In my reading it seems pharma especially is guilty of such things as p-hacking by adding/dropping "problematic" subjects, failing to publish findings for all the original hypotheses, playing games with absolute vs. relative effectiveness and risk, redefining side-effects midway thru etc. many of those slights of hand are in the detail records as I understand it but often not published or available to the FDA. I no longer trust this system, from A to Z. Love PHil's efforts however.
Furthermore, oftentimes the abstract directly contradicts the actual findings of the study.
And I guess an AI could immediately detect when a meta-analysis does not use the proper results from the included studies.
Was I thinking about the Roman et al. meta-analysis on ivermectin? Well, maybe ^^
The GIGO problem ...
That's the problem. Although there is a big 'translation' problem that the author is trying to address. Others are trying to 'open source' the original data and its analysis...just like the FDA tries to do...but the pharma/etc companies will not allow that for 'patent' and 'business' protection reasons. I once joined an effort to set up clinical trials of 'approved' drugs to verify their effectiveness and safety...but that company changed course. That would be a start, if someone could get sufficient funding.
I lack your expertise in this field. But if I understand this article, a basic benefit of AI analysis is the ability to summarize medical studies as either -
#1 - “This data yields widely varying results depending on who funded the study”
VS
#2 - “This data yields mostly consistent results which create a clear course of action”.
Summary #1 creates immediate “public shaming” of compromised studies, and will improve accuracy through peer pressure.
Summary #2 tells doctors and patients they are at least on the right path and not randomly guessing.
I’m excited for ANYTHING that looks like a path forward
well, there's also the wonderful gift this is to college students. this takes cheating to steroidal levels. where was this when I was a student?! /sarc
This is incredible. Literally the best Substack post I've read in 3yrs. THANK YOU, Phil! You're doing God's work for sure.
Am reposting and will highlight your work on an upcoming Substack post of my own, about moving forward with a solution oriented focus.
Bringing a scientific paper down to understandable levels (for your "average" person) is a great goal. My questions are-- will AI be able to objectively tell the reader if the study has any basic inbred flaws in its methodology, are the conclusion drawn in sync with the facts in the study body and are any of the authors influenced (financially or otherwise) by Big Pharma, government agencies, etc. Reading, interpreting, understanding these papers are no small feat. Ask somebody as knowledgeable as Pierre Kory.
This is a great idea. I just hope it will work as advertised. The ability to get these studies out to average people at an understandable level is very important to allowing people to protect and guide their own medical treatments. Thanks for your efforts, Phil.
As someone with one of those pesky autoimmune problems for which no one can identify the culprit, I have been studying “medical studies, journals, etc...” for at least 15 years and the biggest problem I always encounter is that none of us live in a bubble. I completely understand that we need to do studies for the larger population, but until we admit there are just so many other factors many are bringing to the study, nothing will really be helpful. In my case the research and understanding on the ankylosing spondylitis I have had for 40+ years is ever evolving and at the same time quite stagnant. (As an example, so many of the “other symptoms” of AS which happen in less than 20% of those of us with the disease have happened to me while I have no family history, I have never been overweight, never smoked, am healthy otherwise and yet I have experienced these “rare” percentage other issues - everyone of them??? Probability speaks against this) The data input will always be missing very important information that can never really be extracted on a large scale. We must listen to our own bodies as everyone is different and if we pay attention, what we can’t heal through life changes as we attempt to understand the broken pathways, we can try to mitigate. And really, feeling like you are doing something helps mentally which helps physically as well.
while its nice to have a lay translation... how can we know if its reliable in the first place?reading cliff notes on a false narrative doesnt help.. great idea tho
Not inject any "vaccine" into our body, sounds good.
Thank you for creating this tool. It's much needed in this war for truth and for our freedom from those who would lord over conclusions.
I wonder if we could work together. I have built an evidence sharing platform in which summaries like these are organized into a tree of empirical inquiry with a yes/no question at the root. Our first debate was about whether masks mitigate the risk of COVID-19, and contains an amalgamation of evidence on both sides sorted by user-weighed strength, each open to rebuttal. My goal is to create a space where controversial questions can be examined and where all the evidence is visible, understandable, and tested, with no suppression of either side.
I've hired writers to help me summarize evidence and this has been fruitful both for creating new content and for shaping how the inquiries are organized. However, this is slow and expensive, eating into my savings, and probably won't scale. AI like yours can provide an enormous benefit by allowing even my end-users to request summaries of evidence they themselves submit. This puts the power of knowledge and argument back into the hands of society at large and away from the 'speakers of science.'
Would you like me to share a link to this project?
Could we employ AI or ML to help you build / parse your evidence trees? I like the idea.
I think AI will play an important role in finding and summarizing evidence. The parsing of the tree is something I'd still like to leave to the individual. My philosophy for Evincer is to give humans easy access to all the best evidence so that they can decide what's true.
I'm starting to use AI now to find evidence and another AI tool to summarize it. Here's a post I made this way yesterday for example: https://evincer.org/evidence/2009-regression-based-test-shows-no-recent-c-02-correlation
Evincer is an open platform and users often just submit links without summarizing for the reader. So I would like an API that Evincer could access that would help users easily flesh out the missing details. It would accept a link and a debate title: (e.g. "Did Human GHG Emissions Cause Global Warming?") and reply with how the paper addresses the question.
What's the roadmap look like for your project? btw I haven't been able to connect lately. Got a 504 today.
I haven't been able to access the openpaper.sciemce either - I'm not sure what Phil is doing with it, nor how it's being hosted. I also am getting the 504 error. I will go check out your Evincer post - but first skim I do think you could partner up your platform with Phil's AI summary tool.
Outstanding idea, Phil. Even having a postgraduate science degree it is difficult for me at times getting through a lot of advanced science papers.
😱 I did ask ChatGPT, in one session with it, to write a 200 word essay on George Washington at a 7th grade level. It complied very well. AI, like so many advances in science and technology will be used for good and evil. I suspect homework assignments of the future will not be as educational for the student as they once were.
Oh, welcome back Phil!
Love it, keep digging, keep writing :)
Stellar, as always! Excited to see where this goes.
Bravo! Great idea!
Thankyou, love this approach.
I had found all the evidence of corruption you mention in the beginning of your article and more, it's sooo bad. As a retired RN/RM I can understand the gist of some papers but cannot understand the intricacies. Your tool will be invaluable to many. The interest in looking at scientific papers has grown over the last 3 crazy years. I think your tool will be popular.
Well done, Phil, add this to your references? Keep safe and free.
Episode 353 - The Crisis of Science Corbett Report 02/23/2019
In recent years, the public has gradually discovered that there is a crisis in science. But what is the problem? And how bad is it, really? Today on The Corbett Report we shine a spotlight on the series of interrelated crises that are exposing the way institutional science is practiced today, and what it means for an increasingly science-dependent society.
https://www.corbettreport.com/sciencecrisis/
9:30 pm Eastern Standard Time when I tried to register got an error message. I’ll be patient. 😁 Looking forward to it. I live in the medical literature in my “day job.” I’d like to believe that most of my last 20+ years of research has had honest and transparent findings (some of which have been in those 2 big name journals you referenced). Journals don’t like to publish lots of clinical trials that failed. But with requirements to post all study findings within 12 months of completion, you should have more data for AI to learn.
Hi Phil, very excited about your work, I just tried to register for openpaper.science, and it is not letting me register, wondering if this problem will be fixed soon.
Fix incoming.
It's Feb 15 & I'm getting a whirlygig forever. What abt showing a "Under Construction" msg for the link openpaper.science?