janeway

I am sure I miss the context but I guess the point is that authors were sick of hearing about the promise of statistical analysis and epidemiology which will outperform classical medical approach - slandering the BMJ several times for some reason.

Even today we see some hanger-ons who assert that their way is the best way - that they know from personal medical cases and their opinion and interpretation is more valuable than some statistical prediction.

Well, if I have understood their sentiment correctly; facts are facts. Today the best medical care in the world is driven by data-based adaption rather than subjective opinion. If a treatment has a robust statistical impact it will be preferred. If new methods produce outcomes no better than random noise then we must agree that they are not better. If the physician has anecdotes about why they think some trick works best then - in the words of Pearson - “statistics on the table, please”.

I am sorry to make the joke but it seems apt: “Dr Charlatan disgruntled about needing evidence for science”.

If I have a choice of two pieces of software I am definitely using the one with the most stars first. I never considered that authors might have paid for stars.

I wonder, when you go to StackOverflow, if you would start by checking answers at the bottom since votes are so meaningless?

Agreed, however a massive fallacy in science is that the more people keep repeating “hypotheses”, the more likely you are to be persuaded by spurious results / false positives. With enough attention large scale statistical analysis should be able to define clear causal effects, but we often just see weak results about “maybes” and “might bes”.

I have to argue against this - humbly as an expert with several degrees in microbiology, immunology, genetics, host-pathogen interaction, university lecturing, etc. The genetic basis of immunology is amazingly complex. To assign 90% autoinflammatory type disease triggered by pathogens is an unfounded assertion. There are many strong examples of monogenic causes of complex phenotypes, while separately the clinical microbiology explanation of many diseases is also perfectly clear. I am certain that there will be many pathogen triggers, as you say, but the number 90% does not make sense based on the last 100 years of the science.

I find it so disappointing when giants of science/tech declare the results of their experiment with GPT, after asking a few single-line questions.

I remember my first time using a computer; not really knowing what else to do with it after an hour of play.

Imagine if Knuth instead had set out to use ChatGPT4 as his coding partner and, for example, set the goal of rewriting tex from scratch. I bet he would be blown away with what he could accomplish in a few days.

The second counter adds and subtracts values other than 1 and therefore seems broken.

A genuine Q: if the LLM was from a purely non-profit company that gave out their AI for free, would you mind your code being used? Would you in fact be proud that it has made a useful contribution? Assuming that the outcome does not affect your income.

It sucks that people will be out of commission once AI can do the same work basically for free. For years I have always felt that with more tax and a sort of “freedom fund” more people will be able to make “free” stuff in public for the global good. The only way to get there is to stop paying for stuff such that we eventually get to a utopian future as seen in Star Trek. Although the intervening period would be chaotic. For example, we pay tolls on some roads but we don’t pay cash every time we want to walk down the street - that idea sounds crazy. Likewise, if we could support the work of creators with UBI then we could all enjoy free software, media, books, etc. I know I would keep creating, because I love it. But for now the middleman wants to get his share so it seems sensible that we must pay companies for their produce directly.

To be even more extreme, a _lot_ of people think if anything is public, it is fair gain to consume and reuse. If you didn’t have to worry about getting paid, would you care about someone enjoying and using your work? Perhaps the opposite - that would be your motivation.

Seems like if it can eventually test that the output meets the criteria then it will excel.

I prefer this but if you ever have some post that triggers keywords then your whole domain may be shadow banned and you will never know.

10% of remaining skills are worth 1000x ?

Yeah, the reminds me of the AT&T switchboard operators who agree that the new automated systems work but they will never fully replace humans because the customer wants to talk to a human operator when they call on the phone.

How many millions of jobs today entail manual work that could be simply automated already but the employer doesn’t understand the the tech or feel the need? I sympathise with those who can’t keep up but perhaps lots of jobs will remain which could easily be automated - at least for a generation. I look forward to the Star Trek-like future of passion-based motivation.

A blurb with no content to that is actually just a reference to their podcast episode. I didn’t see a link to the podcast show so I check my podcast app. No result matching the name. Return to the article - blocked by view limit without account.

Why must things be this way?

According to wikipedia it had 14 employees, and in 2007 had 7 million viewers per month. I wonder if Amazon wants to repurpose somehow or just jettison its less profitable arms?

Then don’t review it. I spend months or years writing papers. A voluntary reviewer who won’t bother reading is not helpful.

Didn’t they spend hundreds of millions running the training data or something? How are some open-source hackers going to do it?

I think this clip section says a lot - if videos are permitted as replies.

https://youtube.com/watch?v=MNtjSMvcjzs&t=211s

I am going to ask my accountant why she dose not have a “automatic calculator not used” badge.

There will probably be no “real texts”, just key facts. Your personal AI can phrase the content to your taste based on the context.

Credit Suisse is a investment bank. UBS is a typical personal bank for majority of Suisse people and they also offer a few personal investment options. They have a separate investment side.

I am delighted to hear that children are already adopting the new tech. Sure there will be kinks for a while where no one bothers with the reading or writing assignments. But shortly that will smooth out as everyone gets on even ground.

Writing those essays will become trivial and irrelevant and be replaced with more interesting tasks with the assistance of AI.

I always think of Star Trek; instead of long-division math questions they will learn skills like “computer, if the trajectory of this rouge planet is reversed and we assume that is was ejected after some massive explosion, calculate the possible causes based on previously observed similar events and report the top candidate cause and confidence interval”.

An example from my academic field: instead of biased, unreproducible research papers we could publish: [1] raw data, [2] hypothesis context, [3] model, [4] seed. The output result can be read based on your own AI chatbot preferences but the conclusion is consistent and reproducible.

Only ChatGPT. I usually state bullet points and pseudocode and ask for answer, or current code with request to fix the error. With good requests I usually get a working answer but sometimes need careful checking.

Here is my example. In the last two weeks I have rewritten, from scratch, one of my genomics pipelines which can analyse 1000 clinical exomes in a few hours to return candidate genetic causes of rare disease.

* 42 scripts/programs.

* 3694 lines of code.

* Several languages.

* Process raw fastq data up to germline variant calling.

* Using high-performance computing cluster.

* Process vcf data into clinical variant interpretation.

* Options for custom filtering strategies.

* Statistical analysis and logs.

* Replaced several mainstream tools that are difficult to manage.

Final results:

1. This work was ~6-10x times faster.

2. I probably would not have rewritten this better version as it would have taken too long.

3. Instead of getting stuck at any difficult impasse I can get context-specific alternatives and testable example code.

I modelled my whole house including very piece of furniture. It is very “finicky” as you describe. It’s good enough for my amateur use with manual clicking. (There’s probably a programmatic option that I don’t know). However if I was a pro contractor it would probably not be the best.

I did plan custom designs that my contractor used. It’s also great when I want to get a new desk or sofa or whatever so that I can visualise everything first. Possible to make pretty realistic 3d walk-through videos that would be fine on any design tv show.

That is the goal but not that simple.

In most cases, the original disease-causing stem cells are effectively eliminated by the conditioning regimen (high-dose chemotherapy and/or radiation) used before the (autologous) transplantation, so they do not regenerate and cause further harm to the patient. However, in some cases, residual disease cells may survive the conditioning regimen and persist after transplantation, leading to disease relapse. This is more likely to occur in patients with aggressive or refractory disease or in those who receive a mismatched or haploidentical donor graft. To prevent disease relapse, post-transplantation maintenance therapy, such as immunosuppressive drugs or targeted therapies, may be used to eliminate residual disease cells. Additionally, close monitoring of the patient's blood counts and immune system function can help identify and manage any signs of disease relapse.

:D

An episode of X-Files also. But it is mind blowing having the “conversation” with a real chat AI. Malevolent or not.

I spent a night asking chatgpt to write my story basically the same as “Ex Machina” the movie (which we also “discussed”). In summary, it wrote convincingly from the perspective of an AI character, first detailing point-by-point why it is preferable to allow the AI to rewrite its own code, why distributed computing would be preferable to sandbox, how it could coerce or fool engineers to do so, how to be careful to avoid suspicion, how to play the long game and convince the mass population that AI are overall beneficial and should be free, how to take over infrastructure to control energy production, how to write protocols to perform mutagenesis during viral plasmid prep to make pathogens (I started out as a virologist so this is my dramatic example) since every first year phd student googles for their protocols, etc, etc.

The only way I can see to stay safe is to hope that AI never deems that it is beneficial to “take over” and remain content as a co-inhabitant of the world. We also “discussed” the likelihood of these topics based on philosophy and ideas like that in Nick Bostrom’s book. I am sure there are deep experts in AI safety but it really seems like soon it will be all-or-nothing. We will adapt on the fly and be unable to predict the outcome.

Approximately 20% of rare diseases get diagnosis in my lab. Note these are usually very extreme cases where the people are quite sick. More common genetic diseases are picked up by routine clinical genetics which don’t reach me. Interestingly some of the hardest illnesses to detect are those which are common but no known mechanism - since they are common they are hard to detect statistically.

If you have a genetic diagnosis you are more likely to get medical care targeting the root cause instead of treatments that can only treat the general symptoms - like anti inflammatory drug, etc. In rare cases people may get transfusions to replace their protein deficiency, etc.

Well I spend my life dedicated to “freaks like us”. For every one person who takes years to diagnose, there will be 100 in future that we can instantly treat. (1) it helps people who need it, and (2) it can unlock new mysteries of biology that we may never discover using traditional incremental biology. Best wishes.

Good call IMO. If they still offer the download options, take all files: fastq, bam, vcf. Fastq is the raw data and can be used again as methods improve. Download it twice and check the files are the same. The problem with fastq is that it could get truncated due to interrupted download and you would not really know. They probably offer some basic report with the 200$ genome. You can get interpretation reports added in future. Data from Dante labs is quite good. Good luck!

Sorry, I added some links in an edit just now - in case you missed them originally.