Rank: #10 out of 4859 in peer accuracy at Metaculus for the time period of 2016-2020.
ChristianKl
Karma: 35,864
I agree that commercial models don’t detail their data, the point is to have an estimate.
That’s I searched the key’s under the streetlight. The keys are not under the streetlight.
I guess, Soldaini et al., ‘Dolma’, made their best to collect the data, and we can assume commercial models have similar sources.
Soldaini et al have far less capital to collect data than the big companies building models. On the other hand the big model companies can pay billions for their data. This means that they can license data sources that Soldaini et al can’t. It also means that they can spend a lot of capital on synthetic data.
Soldaini et al does not include libgen/Anna’s Archive but it’s likely that all of the big companies besides Google that has their own scans of all books that they use do. Antrophic paid out over a billion in the settlement for that copyright violation.
Even outside of paying for data and just using pirated data, the big companies have a lot of usage data. The most common example for syncopancy in AI models is that it’s due to the models optimizing for users clicking thumbs-up.
A key aspect of modern democracy with the rule of law is that companies can operate even if people believe they are acting with bad character. It’s not hard to convince a majority that Elon Musk and Sam Altman are people with bad character but that’s not sufficient to stopping them from building AGI.
As far as “should have freedom of speech and press” goes, both Republican and Democratic administrations over the last two decades did a lot to reduce those freedoms but the pushback comes mostly on partisan lines. They amount of people who take a principled stand on freedom of speech no matter whether it’s speech by friends or foes is small.
As far as “should have a monarch who inherited legitimately” goes, I think it worked for a long time as a Schelling point around with people could coordinate and not because most people found the concept of being ruled by a king that great. It was a Schelling point that allowed peaceful transition of power after a king died where otherwise there would have been more conflict about succession.Eg JK Rowling’s character Sirius’s claim that you can see the measure of a person by how they treat their house-elves
While we are at general principles, citing JK Rowling in a discussion on ethics is probably generally a bad idea for politics is the mind killer reasons. I think the article is very interesting in terms of cultural norms.
It gets frequently cited to make a point that discussing politics is inherently bad, which isn’t something the article argues. On the other hand, the actual argument that if you use political examples it will make your audience focus on politics and make them less clear thinking when you could use non-political examples that don’t have this problem is seldomly appreciated, because people like using their political examples.
Proof-of-stake is just technology that can be used in different ways. It can be used for pump and dump scams but also for different purposes.
If you are building a product that’s actually setup to create long-term value it’s useful to use proof-of-stake is it allows you to provide more value because you have higher throughput while using less energy.
If the value of the project rises as features are build out, there’s an incentive to build out the project. There’s no good reason for anyone who wants to build a system that actually creates value to use technology that burns more energy and provides less performance.
We train LLMs not only on the artifacts from our best thinkers but, in 99.95% of cases, on web crawls, social media, and code.
Concluding from a paper that says in it’s abstract “commercial models rarely detail their data” that you know what makes up 99.95% of cases of training data, is a huge reasoning mistake.
Given public communication it’s also pretty clear that synthetic data is more than 0.05% of the data. Elon Musk already speaks about training a model that’s 100% synthetic data.
From the perspective of thinking that Tyler is innocent it makes a lot of sense to encourage him to turn himself in. It makes it less likely that he would get shoot during the arrest and also makes it less likely that someone else can shoot him in retaliation for allegedly shooting Charlie. He shouldn’t have any expectations of permanently being able to escape a manhunt either.
B) Sing from the rooftops and to every single possible news outlet you can find that your son is being setup in order to free him.
Do you think that’s what a lawyer would advise him to do as the most effective way to free his son? I doubt that’s the case. Standard legal advice is to not to talk to the media.
But this solution should not be appealing to anyone who wants to use the cryptocurrency even if a cryptocurrency is better funded without much mining (of course, if mining is replaced with another consensus mechanism after all the coins have been created, then this objection does not stand). After all, Satoshi Nakamoto did not fund Bitcoin by selling bitcoins. There are other ways to fund a cryptocurrency project without alternate consensus mechanisms.
I don’t understand why that would be an argument against just using proof of stake. Proof of stake has a bunch of different benefits. It solves the energy problem.
It also increases the amount of writes that the blockchain can do per minute which is very important for usability.
How about reading the post to which I’m replying? It quite explicitly defines bad faith in a way that not about “conscious intent to deceive”.
If you look at the YIMBY example that Anna laid out, cities policies are not under direct control of citizens, yet Anna found some points that relate to what people can actually do.
If it seems like you don’t have any control over something you want to change it makes sense to think of a theory of changes according to which you have control.
Right now, one issue seems to be that most people don’t really have it as part of their world view that there’s a good change of human extinction via AI. You could build a heuristic around, being open about the fact that there’s a good chance of human extinction via AI with everyone you meet.
There are probably also many other heuristics you could think of about what people should do.
For whom would you recommend getting additional doses of the Covid vaccine right now?
The scenarios are either that Tyler Robinson is the perpetrator or that someone tried to set him up as a patsy.
In an age of no-click remote exploits for phones, it’s easy for someone with nation state capability to fake the text messages and there is plenty of the discussion online that they are looking fake. I think it’s debatable whether the text messages are stronger evidence for Tyler as perpetrator or patsy.
The rifle being found in the woods near the shooting is consistent with both versions.
confession
It’s a hand-written document that could also have been faked if someone tried to setup Tyler as a patsy. The same goes for the discord message.
convinced many civilians, including Robinson’s own father and his boyfriend to lie
We don’t have on the record statements from Tyler’s father. We have statements from the police that suggests that the father thought Tyler is guilty. There are statements from Candace Owens that she contacted Tyler’s parents and that the position of his parents is that Tyler is innocent.
The official position is that the boyfriend (Lance Twiggs) has no prior knowledge of the killing. To the extend that he knows anything about it, it’s that he received text messages. There’s also no on-the-record statement from Lance that Tyler is guilty and it seems people who want to interview him can’t reach him.
The TPUSA position seems pretty weird to me. Their [spokesperson Andrew Kolvet is essentially saying](https://www.youtube.com/watch?v=6VzyGCRB4B8&t) that he privately asked some people in the Trump administration to look into Israel as a possible suspect for the killing. That they are glad, that Candence publishing messages put them into a situations where they could say so, because apparently beforehand they didn’t think they could be open about that.
They also seem to argue that it would be harder to convince a jury that Tyler is guilty if TPUSA would release more of the video material of what went on that day, and that’s why TPUSA doesn’t release more of the video of that day.
I somehow had a cached thought from looking at Metaculus statistics of myself in the past that my performance was more middle of the pack. Yesterday, I looked at the statistics again and it ranks me, Rank: #10 out of 4859 out of Time Period:2016 − 2020.
I’m not exactly sure how I got my impression in the past, maybe I was better at the more long running predictions? Or did Metaculus change something about how it calculates scores?
“Learned helplessness” is a quasi-popular-therapy word for this.
It’s a terms that means something else:Learned helplessness is a psychological state in which a person (or animal) stops trying to change their situation after repeated experiences of powerlessness, even when escape or improvement later becomes possible. The concept originated from experiments by Martin Seligman and Steven Maier in the late 1960s: dogs exposed to unavoidable shocks later failed to avoid shocks they could escape.
In the situation the OP talks about it’s possible to change the situation by signaling being in pain to other people and then getting help from those people or those people otherwise accepting behavior of the person to change the situation that they might not otherwise expect.
I’m not entirely sure how this relates to seeing this pattern in others. If they’re your friend, you typically take a two-pronged approach: help superficially and get them to address the root problem to the extent you’re aware of a root problem.
Who’s that “you”? There’s stereotypical situation where a wife tells her husband about one of her struggles and then the husband tries to superficially help which actually doesn’t lead anywhere.
I think a good default response is listening and holding space for a person that suffers. There are also higher skill options that involve not accepting the frame. If you want to understand more in that regard jimmy’s sequence is good.
I like geeking out over masks and there are a lot of options. I have a bunch of models, and if you’d like to come try them sometime (next EA Boston meetup on 10/26?) I’d be happy to show you what they’re like.
I think it would be great if you can gather some data about how many people prefer which of the options and publish it.
Let’s say you have a leader of a company that uses AI a lot. They make some decisions based on the advice of the AI. People who don’t like those decisions say that the leader suffers from AI psychosis. That’s probably a scenario that plays out in many workplaces and government departments.
I’m a good prompt engineer
You are vibe coding
He has AI psychosis
A new major study finds that alcohol causes cancer, so government worked to bury the study. Time and time again we get presented with the fact that small amounts of drinking correlate with improved health in various ways, fooling many into thinking a little alcohol is healthy.
That seems a pretty uncritical way to frame the issue.
It turns out when you put a few people who believe that nutritional research with it’s correlational observational studies is crap into leadership positions, they don’t take that kind of nutritional research very seriously. MAHA was never about trusting the existing nutrition researchers, so this should not be surprising.
If you do report on the issue, I think it would make sense to focus on the actual merits of a policy choice instead of just “Trump administration doesn’t like the status quo that nutritional experts propagate—Nutritional experts are so worthy of respect that disrespecting them is bad”.
It’s visible in several frames as he walks away, otherwise it blends in with his legs.
What time stamps do you mean?
Which blob are you talking about?
While the moment right before he jumps might have a blob that’s consistent with a lot of different items, it seems to me like the time he walks, there’s no such blob as far as I (and GPT-5 Pro) seem tobe able to tell.It’s worth noting that the government document says that him carrying an object consistent with being a rifle is visible while he runs across the roof. The moment where he prepares his jump is not a moment he runs over the roof.
I have not told it anything about me not seeing the rifle. I did give the link to my chatlog and the questions I asked, and my follow-up questions. In a previous chat I just asked it fairly objective questions about what the official timeline is. Given the way GPT is set up, if anything that should bias it toward validating the official version.
I don’t think GPT-5 Pro’s capability is absolute proof, but it’s another set of eyes and GPT-5 likely know a bunch about how artifacts from low resolution in videos are supposed to look like that I don’t know.
If you create a system based on your own experimentation in a psychological field where you deviate from what’s normal and you crate a bunch of terms to have a handle on what you are doing, you should assume that those handles are unique to yourself.
There might be someone who does something similar than you, but they are likely not using the same vocabulary.
If you want to ground yourself, journaling frequently about the process is a good idea. It’s also good to expose yourself to real world feedback.
What would you say about Tucker Carlson (while he was at Fox) and Rachel Maddow (MSBC) who argued in court that their audience doesn’t take them literally and know that they are just entertainers as a defense against claims of defamation.
I would expect Zack to say that both engage in Bad Faith when they are saying things that are not true but work for engaging the audience.
While the audience knows they are watching partly because of entertainment, the overt appearance of any statement is one of being told the truth. The same goes for questions asked during a friendly conversation. We know we are together because we care about spending time with each other but the surface appearence of any question is still about wanting to know the answer to that question.