How could consistently positive results with large effect sizes persist for 30 years if the peptide is truly ineffective?

A majority of research studies in homeopathy find clinical effects. If you however limited yourself to high quality research papers, homeopathy seems to provide no clinical benefits. There are plenty of anecdotal reports of people getting large effect sizes from homeopathy. 

It's the key argument for evidence-based medicine. Various placebo effects frequently make people believe that they have effective treatments when the treatment itself does nothing. While I argued that you could switch to prediction-based medicine and have an alternative to evidence-based medicine, medicine by anecdotes might just be a bad paradigm.

How do we explain the independent researchers in Taiwan/China/Korea? Are they complicit or just publishing invalid results?

Any honest researcher, that publishes research showing that BCP-157 is great, should be interested in what the protein actually does in the normal metabolism. Not doing the literature review to find out that BCP-157 does not seem to come from an exciting protein suggests complicity.

Publishing negative results is generally hard and there's little checks to scientific fraud in the far east.

Why establish a company and pursue patents if it's entirely fraudulent?

As long as people are willing to buy a treatment the company makes money even if the patients don't get results. BCP-157 is popular. 

How plausible is the existence of an undiscovered stomach protein?

We sequenced the genome of most mammalian species, so there are no unknown mammalian proteins. To have an unknown undiscovered stomach protein, it would need to be produced by a bacteria that lives in the stomach that nobody has identified and sequenced. That seems implausible in 2025 given that we have shotgun sequencing to identify all bacteria in a given area. Remember that the bacteria would need to produce enough of the protein for the protein to be harvested the way Predrag claimed. 

How has no one in the fitness/biohacking community noticed it's completely bunk?

That does suggest that there's a problem with those communities and people in general not doing deep research. Even when I emailed the people at Examine.com, that wasn't enough to get them to update their article to show that BCP-157 is likely complete bunk. The communities seem to be trend-driven. Influences make a lot of money with affiliate deals.

I'm at 10%, 25%, 50%, 15%. What about you @ChristianKl?

When you inject a random peptide, you get an inflammation response in the area where you injected it. There might be cases where that produces a useful therapeutic effect. 

Apart from that I would give complete fabrication maybe 95%. 

I focused my post on the argument that the BCP-157 doesn't come from a know protein, but that isn't the only argument to be made.

If Big Pharma companies would believe that BCP-157 is real, they would likely want to produce drugs that target the mechanisms through which BCP-157 produces it's effects. 

Bryan Krause answer to my Stackexchange post analysis the implausibility of the early BCP-157 research in more detail. 

It's been a while since I looked into the animal research, but from what I remember given the size of BCP-157, it shouldn't be absorbed from the intestines into the blood, but the animal research suggests it has some effects when ingested orally, that would not just require it to have an effect but also be absorbed from the intestines into the blood. That's another sign that the animal studies are crap.

Yes, what's covered by price gouging would likely expand. 

Politicians who want to balance the budget will find it easier to argue to expand the revenue through expanding what's covered under price gouging than to raise income or sales taxes. Opposition to the proposal would also be harder than to oppose what's currently covered as evil socialist price setting by the government. 

Kamala campaigned on making more price gouging illegal than currently is illegal. Thinking that this will only ever apply to the type of price gauging that was previously illegal ignores how the politics are likely to play out.

There are a variety of clever way to get something by making tax law more complicated. In general, I think we aren't skeptical enough of increased bureaucracy of the tax law and as a result our tax laws are way too complicated.

Whenever one advocates a way to make tax law even more complicated, I think it's important to be explicit about the cost of the increase in tax law complexity. 

Having with Cerebrolysin and BPC-157 the two top-rated peptides to be bogus, does suggest that the whole field is untrustworthy. It also makes me more skeptical about self-reporting.

That's not a good data point. If you want to provide anecdotal data, it would be good to provide more of the observations. How long did he have a shoulder issue before taking BPC-157? How fast did it get away afterward?

What dose do you believe to be good for that?

It's not just a question of whether people agree but whether they actually comply with it. People agree to all sorts of things but then do something else. 

When it comes to blind spots, we do have areas like medicine where we don't pay as a result of outcomes of medical treatment. That leads to silly things that when surgeons say that having 4k monitors will obviously improve the way they do surgery because it allows them to see details that they otherwise wouldn't, without anyone running a clinical trial that shows 4k monitors to be superior, they don't get adopted. 

Evidence-based medicine is a strong dogma that prevents market economies from making the medical provider that creates the best outcomes win. 

What does it add to the debate?

That sounds like you believe that morality is something that exists in some platonic sense. And that you know what the platonic entity is in a way where you can be confident that slaves have a lower moral weight.

Otherwise you are simply choosing to use your own definition for morality that's quite different from what other people mean with the term 

By that reasoning unwilling human slaves shouldn't get moral consideration either. Is that what you actually believe?

There are mechanisms where fluoride goes directly from the mouth onto the surface of the teeth. There are also mechanisms where fluoride goes from the bloodstream into teeth.

The fluoride that goes directly from the mouth to the surface of the teeth seems clearly good for caries prevention at low side effects.

When it comes to the fluoride that goes through the stomach and blood supply, it's unclear to me whether that provides a benefit for caries prevention when you already have sufficient fluoride through toothpaste in the mouth. The side effects also seem unclear to me.

Given that people eat shrimp, their contribution to human society is not zero.

With Trump speaking about the US making Greenland a US territory, it would be worth speaking publically about how the US screws their territories with the Jones Act.

It might be a way to build momentum to get rid of the Jones Act.

Drug legalisation is probably the best way to prevent fentanyl deaths. Many of the people who are fentanyl addicted are addicted because they wanted to buy other drugs buy got them laced with fentanyl. Drug legalisation will allow for quality control and end the lacing with fentanyl.

Now, nitazene which is even more potent than fentanyl gets added and might produce similar effects as fentanyl where people die to nitazene overdose.

If the US wants to actually prevent those deaths drug legalization that allows for quality control is the step forward that would work.

 

With o1 and o3, we found out that you get better output if you spent more compute on getting the answer. Letting an LLM work 10 minutes via chain of thought on getting an answer gives you better answers than letting it work 2 seconds on getting an answer. 

By training on it's own answers you could get the LLM to give the kind of answers it gives with 10 minutes of chain of thought worth of compute in 2 seconds without chain of thought. 

The difference between a person with an IQ of 90 and one with an IQ of 180 is not in gaps of knowledge or having access to information that's right or wrong but in reasoning ability. 

As far as the territory of LLMs being "out there", that's true for a subset of the territory. The field of mathematics isn't "out there". You don't need to run any experiments in physical reality to validate claims about math and see whether you get better at creating math proofs. 

When it comes to computer programming, some tasks require interaction with physical reality but plenty of tasks are "write a patch that includes a functionality that passes all the unit tests and passes some clean code metric". You also have "find a possible input for a function that makes it crash" and a few other tasks where you can automatically evaluate the outcome of what the LLM is doing without needing any human input. 

Markets let us buy anything society can produce, anytime.[1] At first glance, that sounds great.

While that might be true for physical goods, it's not for other goods like friendship that people care about.

It's a combination of the way LLM work that they predict the most likely token that very similar to prediction common wisdom and experience of interacting with LLMs.

Pattern matching also matters. After reading the answers from Claude and ChatGPT, you can ask yourself what you expect a person to tell you when you ask them for the top five reason and how likely it is that they will tell you "sports online betting" as one of the top five reasons. 

The key problem with pay-on-results is that it creates disincentives for clients to achieve their goals. If someone invests money into a coaching, that's a commitment device to work on achieving results in the coaching. 

I know a few people who were doing hypnosis for treating allergies. One friend who let themselves be paid succeeded in three out of three cases. Another person wanted to do a study and recruited patients for free study participation. It mostly didn't work with the people who didn't pay. 

I wrote about prediction-based medicine as a system that both creates buy-in for the patient who's committed while having the same similar incentives for treatment providers to solve people problems in a minimum amount of time. 

A priest or warrior might imagine that they are possessed by a god when lying to peasants or murdering enemies, but possessed by some demon when seeking forbidden intimacy or abandoning a fight. 

Buddhist have some norm against forbidden intimacy. The way to get around them in tantra is to identify with a God while having sex and not act with the normal human identity.

Jewish tradition has similar methods where the way to get around the sinful nature of sex is about identifying with god during the act. 

As far as I remember the RadVaC people were working on mRNA as well. If you want to do something like this, their Discord is likely a good place to check and ask questions.

If it would be only true in the case of calorie restriction, why don't we have better studies about the effects of salt?

People like to eat together with other people. They go together to restaurants to eat shared meals. They have family dinners. 

The main problem of nutritional research is that it's hard to get people to eat controlled diets. I don't think the key problem is about sourcing ingredients. 

Egyptians felling all their trees and turning their environment into a desert feels quite similar to fossil fuels. 

A rationalist interjects: “You should make public predictions about this stuff!” Idk, should I? What should I make predictions about? About whether individual cases succeed, or some broader trends? I'm not sure if it’s worth my time. I really like $ as a metric, not sure what the predictions add. Very open to being convinced here!

Predictions about individual cases would be great. Whenever you take a deposit write down the condition for the bounty being paid out, the amount of the bounty, and your self-assessed likelihood of the person paying the bounty in the following twelve months to you into a public Google Sheet. Maybe, add another row for "time-spent with the person".

The exercise about thinking beforehand about how likely you will solve the issue for the person is useful for you to understand your method better. It also help informing potential customers well about what they can expect from your service. 

Finally, it would be great to have a one-year follow-up after a bounty is paid and that information also added to the Google Sheet.

Anthropic should have a clear policy about exceptions they make to their terms of use that includes them publically releasing a list of each expectation they make for their terms of use.

The should have mechanisms to catch API users who try to use Antrophics models in a violation of the terms of use. This includes having contracts that allow them to make sure that classified programs don't violate the agreed upon terms of use for the models. 

Which one's do you see as the top ones?

That sounds like it's relatively easy to game by the company who chooses the investigators.

Exploitation is using a superior negotiating position to inflict great costs on someone else, at small cost to yourself.

I think the word exploitation as it's generally used, is about one party getting a benefit at the expense of another party. It's not about one party getting nothing/pays a small cost while the other party suffers a lot.

Promoting an alternative definition of what it means to exploit is likely going to make reasoning harder. Google suggests as definition for exploit "make use of (a situation) in a way considered unfair or underhand". 

Wage theft is a clear example of exploitation. For many jobs, there's information asymmetry where the person seeking the job does not get informed fully about how his job will be before they accept the job, that's also clearly exploitation. Multiple-level marketing companies like Amway are exploitative because they mislead people about the likely results of working for them.

In general, there's value created through trade. If one party captures nearly all of the surplus value of the trade, many people consider that unfair and thus exploitative. 

A key aspect of your examples is further that total utility might not be maximized and because one party has little power, utility maximizing trades don't happen. That's a different issue from how the trade surplus is distributed.

If people complain about Amazon, to my knowledge most of the people complain that while Amazon runs very efficient and is run to maximize total utility, they capture most of the generated value and don´t pay their employees very much. 

Maybe, economists do have a term for the case where one party being powerless leads to utility not being maximized?

I think it would be good to automate the moderation process. Current LLM should be able to make the decision about whether a post is containing the kind of profanity that would lead to account bans.

Annoying civilisational inadequacy:

USB-C cables differ a lot. Some only allow power delivery and no data, while others support different levels of data transfer. Power delivery capabilities also differ.

Most cables do have an E-Marker chips that contain the relevant information. However, Android does not provide that information to the user when they plug into an USB-C cable. 

Edit: After looking more into it, it seems while some cables do have E-Marker chips, most don't :(

Is answer assumes that you either have a fully chat based version or one that operates fully autonomous.

You could build something in the middle where every step of the agent gets presented to a human who can press next or correct the agent. An agent might even propose multiple ways forward and let the human decide. That then produces the training data for the agent to get better in the future.

You could say that Wikipedia falls into the category but given the way it's discourse goes right now it tries to represent the mainstream view. 

For specific claims, https://skeptics.stackexchange.com/ is great. 

https://www.rootclaim.com/ is another project worth checking out.

You see that each of the project has their own governing philosophy, that gives the investigation a structure. 

Yet discourse about these topics more than anything else fundamentally combats propaganda and misinformation. 

The phrase "combat" is interesting here. Julia Galef speaks about the soldier mindset and the scout mindset. Combating anything is essentially about the soldier mindset. On the other hand you need the scout mindset to think well and come to correct conclusions.

By in large the movement that bills itself as "combating misinformation" is about defending the hegemonic Western elite discourse. It's not about truthseeking. 

When reading posts about AI development I get the impression that many people follow a model where the important variables are the data that, out there in the world, the available compute for model training and the available training algorithm.

I think this underrated the importance of synthetic training data generation.

AlphaStar trained entirely on synthetic data to become much better than humans.

There's an observation that you can't improve a standard LLM much by retraining it by just feeding it random pieces of it's own output.

I think there's a good chance that training on the output on models that can reason like o1 and o3 does allow for improvement.

Just like AlphaStar could make up the necessary training data to become superhuman on its own, it's possible that this is true for the kind of models like o3 simply by throwing compute at them.

Why do you care about how effectively the iron in iron supplements gets absorbed? The iron that's not absorbed just gets flashed out. Can't you just supplement more to get what you need?

It's worth noting that the Californian choice isn't free. Californian like residential solar to allow homeowners to feel good about themselves and use net metering to incentives residential solar. Grid electricity in California are double of what residential customers in Texas pay.

Why do you think it would require a central planner to implement agrivoltaics but the profit seeking market isn't doing it on their own?

Your first post is about optimal policy. The optimal response to usually-bogus-but-impactful objections is permitting reform.

How do you know that if you would get rid of net metering subventions which are about letting other energy produces pay for residential solar and other subventions for residential solar, it would still be economical to build residential solar in the US over specialized installations?

You need to pay anyway for the gas plant if you want to have electricity even on days where the sun isn't shining.

You can optimize for different goals. If you want you could optimize for a minimum of new land use. That would however be stupid economic policy as there's enough land and cheaper energy is more valuable.

Using central planing to enforce more expensive energy production because agrivoltaics are cool and reduce land use is not good policy.

What do you mean with "meaningful action" regarding climate in the 2010s?

If you look at solar energy, panel prices go down largely in a straight (on a logarithmic scale) since 1975 with a short pause between 2005 to 2010. German pro-solar policy started in the 1990 and the biggest change in 2011 was the new five-year plan of th CCP which was probably driven more by economic justifications.

There are few people who currently think we should do whatever is possible to reduce temperature rise. Doing so would mean to have a plan for geoengineering. Climate activist often use climate as a justification to push for anticapitalist policies that they independently believe.

Viruses (including phages) depend on the protein production of the host cell. If the host cell codes proteins differently, the proteins that the virus "wants" to produce don't get produced. 

While we still don't know a lot about phages, I think it's likely that phages are a key part of why we have the biodiversity of bacteria in the ocean and other environments that we have. The more of a given bacteria exist, the easier it is for phages to keep the populations of that bacteria down. 

If someone would take a bacteria from the ocean where the population is currently limited by phages, modifying it to change its coding and thus not express the phages that currently target it might massively increase the population of that bacteria.

It seems to me that it's going to be easier to build a bacteria with changed coding for amino acid then to get a whole mirror organism bacteria to work.

Having a 4-base pairs per amino acid coding where a single mutation does not result in a different amino acid being expressed and is a stop codon is useful for having a stable organism that doesn't mutate and thus people might build it.

You get the same problem of the new bacteria being immune against existing phages but on the plus-side it's not harder for the immune system to deal with it.

Instead of focusing research dollars on antibiotics, I would expect them to be more effectively spend on phage development to be able to create phages that target potentially problematic bacteria.

There's a form of exposure that does lead people to suppress certain emotions and become more functional in certain contexts. I think that frequently happens when people try to do a bunch of exposure therapy. 

Here’s a striking fact: if you give a few thousand dollars to charity, you can save someone’s life.

Your link does not support that "fact". The page speaks about estimations.

In their own FAQ GiveWell says:

"However, estimating the cost to save a life involves simplification, guesswork, and subjective judgment. As a result, our estimates are very uncertain. We analyze the cost-effectiveness of programs primarily because doing so helps us see large, clear differences in good accomplished per dollar spent and because working on the models helps us ensure that we are thinking through as many of the relevant issues as possible. For more on how we use cost-effectiveness estimates in our grantmaking, see this page."

Your argument is that the subjective judgement of people, who you like and who themselves say they are very uncertain, is a striking fact. 

This kind of epistemic laundering is either dishonest or ignorant. 

Sam Bankman Fried was a billionaire inspired by the ideas of effective altruism, who ran a crypto firm.

Sam Bankman-Fried wasn't just "inspired by the ideas of effective altruism". He was one of the main case studies of 80,000 Hours who supposedly coached him into entering finance which he wouldn't have without the help from 80,000 Hours. That was true even before he had a lot of money to donate. 

What EAs look for, before endorsing systemic reform, is genuine evidence of effectiveness. For this reason, EAs tend to think that, say, advocating for communism is unlikely to be very effective. We’re not having communism any time soon, even if it is a good idea (I don’t think it is, of course).

This seems to me to ignore the real dynamics. EA do invest money into longtermist courses that are supposed to produce significant systemic change (an aligned AI would constitute significant systemic change) and the evidence for individual causes like for example huge OpenPhil investment in OpenAI is quite thin. 

Given the relative lack of cybersecurity, I think there's a good chance of LessWrong being hacked by outside parties and privacy be breached. Message content that's really sensitive like sharing AI safety related secrets likely shouldn't flow through LessWrong private messages. 

One class where people might really want privacy is around reporting abuses by other people. If Alice writes a post about how Bob abused her, Carol might want to write Alice a messages about Bob abusing her as well while caring about privacy because Carol fears retaliation. 

I think it would be worth having an explicit policy about how such information is handled, but looking at the DM metadata seems to me like it wouldn't cause huge problems. 

On NEPA, the recommended reforms seems to be bills to be passed and not just executive orders. Given that it will be a lot easier for DOGE, to propose executive order than to get bills through the Congress and the Senate, it would likely be good if the process policy agenda would also have suggestions about executive orders to pass to improve the situation. 

This basically sounds like there are people in DC who listen to the AI safety community and told Andreesen that they plan to follow at least some demands of the AI safety folks.

OpenAI likely lobbied for it.

The military people who know that some physics was classified likely don't know the exact physics that were classified. While I would like more information I would not take this as evidence for much.

What evidence do you have for the claim that major Democratic party insiders counseled him to stay?

Dean Philipps didn't win. I think Cenk Uygar got defunded.

If somebody does not pick a fight that's costly to them, that's no sign of careless thinking.