This might be good ground for an adversarial collaboration.
It’s difficult to take a context-less view of hard facts—e.g. rate of non-violent crimes committed per year is heavily dependent on whether the police are bothering to record crimes this year or not, and that’s downstream of a political decision and a cultural impetus.
I’m not sure how a bet could be formulated, and it’s possible that the process of formulating it would show we aren’t in disagreement.
However, I wonder if you are looking at the successful outcomes without considering the unsuccessful outcomes?
A great engineering manager is likely great at engineering. However a great engineer is unlikely to be a great engineering manager—dealing with code all day is different to dealing with people all day. Of course, some can do both (and here you see a great engineering manager who is also a great engineer). My contention is that a minority of engineers (main focus is code) are suited to be engineering managers (main focus is people).
This is essentially the Peter principle. Peter is great at something, and keeps getting promoted until he reaches the level of mediocrity, where he stops getting promoted. Peter is clearly great at position n-1, but not great at position n.
the best managers are chosen to be strong performers in the task that they are managing other people to do.
The opposite is true in most situations, for example a great salespeople will often make a bad sales manager. This is Very Common. Eg https://www.library.hbs.edu/working-knowledge/micromanagers-in-the-making-why-salespeople-struggle-to-lead.
Doing the task and managing the people doing the task are separate skills, and the existence of one doesn’t imply the existence of the other.
I have the opposite experience. It delights me and I enjoy digging in deeper.
People are different!
Despite ceding this ground, I still maintain my fundamental belief that all human beings (and, increasingly I would argue, all life) are worthy of a deep and sincerely-held respect, regardless of any intellectual or physical discrepancies between them. This is not a utilitarian argument, but a deontological or virtue-oriented one, where the basis is one of universal love and compassion. No one should be deprived of a socially and physically healthy standard of living because of the circumstances of their birth or genetics. Nor are people’s wants and needs more or less legitimate due to their intelligence, culture, belief or any other factor.
I want to see if you are consistent. Do you support banning abortion and abolishing the death penalty?
I can just make it look pretty and put commas and periods inside the quotation marks
Maybe it’s from growing up with British English, but I think this looks much worse. Like unclosed or incorrectly closed parentheses.
This acquisition isn’t just about a license; it’s Polymarket’s homecoming, returning stronger and ready to serve American users once again.
Aahh… smell the AI.
6CO2
...6O2?
Suppose these parties do certain things destructive to the country but effective for gaining power, for example, populism.
I don’t understand this, and the use of ‘populism bad’ in general.
I understand populism to consist of appealing to the mass of ordinary people and promising them things they want, in exchange for votes—even if those things run counter to the interests of established institutions and agencies.
How is this different to any other rational action taken by politicians seeking to be elected in a democratic society?
In a democracy, you need votes, and lots of them. People will give you their vote if they feel you will do good things for them. You hear what they say and promise to enact it if you are elected. Other people are unhappy about that, but you do it anyway.
Is that populism, or is it describing every election ever?
I know this is overly simplistic! I don’t understand it enough to steelman ‘populism bad’.
What are the most compelling arguments for ‘populism bad, and populism different to normal political activity’?
Anecdotal, but GPT-5 (mini, I guess? free plan with no thinking) is the first model to succeed at a poetry-based prompt I’ve tested on a lot of models.
I don’t want to mention it publicly, but it involves a fairly complex rhyming scheme and meter.
All other models misunderstand entirely, but GPT-5 got it straight away.
Interestingly, when thinking mode kicked in after a few prompts, it performed a lot worse.
I don’t think the coach analogy is apt. While they may have played the sport, their role is getting the best out of a team of people—a manager, rather than a technical contributor.
A better analogy may be an editor. Many editors are failures as authors, but are very good at critiquing starts, seeing where the flow and pacing needs improvement, and improving the overall work.
However in a world where many editors come to you and submit feedback with varying and contradicting messages, you need to quickly filter by something, so you can focus your limited time and resources on the most valuable submissions.
This is relative to the time and attention that each author has available. Someone with nothing to do will be happier to accept comments than someone who for whatever reasons just doesn’t have time right now to engage.
Prior experience with creating the subject matter may not be the best filter, as you’ve pointed out in the post.
I’m curious what you think might be a better filter for assessing credibility and quality, quickly.
Or do you disagree with the notion that people need a filter?
I hold that — given my experience — I was more justified in my belief than anyone who claims that men playing against women for the World Cup would be unfair. All it takes is trusting that people believe what they say over and over for decades across all of society, and getting all your evidence about reality filtered through those same people. Which is actually not very hard.
So, given this happened—was there any update in your belief in the truthfulness of the other beliefs of those people?
What other embarrassingly unequal parts of reality are being politely ignored, except by science-illiterate jerks?
Excellent review. This is an area that I’ve been thinking about, but don’t know enough about the tech to make a start in a small way.
There’s a company in NZ that takes an interesting approach to solar—Supa, https://www.supa.energy/ (not affiliated in any way).
Ignore all the marketing fluff on the website, how it essentially works is:
Supa approaches companies with large buildings, and gets them to install solar panels + batteries. These are dramatically over-provisioned, they install much more than the company uses
The company buys the solar panels and batteries through a 10-year programme of monthly payments through a finance company, and Supa repays the company monthly for all costs—capital and interest. At the end of 10 years, the company owns the equipment.
Supa arranges and pays for installation, and they manage the energy ongoing. With this, they charge the batteries from the sun (free) and overnight (when prices are lower), and:
Give the company a discount on their electricity
Sell electricity to households nearby, at a discount
Sell back to the grid when the predictable morning and evening spikes occur
The company gets solar, and cheaper electricity with no down payment. Supa gets locations all over the country generating power for them, without having to buy land or build solar farms. The households get cheaper power.
The factors that seem to make this work are (a) NZ’s screwed-up electricity supply—prices are volatile and have risen a lot lately, leading some electricity-intensive businesses to shut down, and (b) cheaper to distribute power generation and storage throughout the country, than to generate in one place and deal with the high costs of upgrading transformers and surrounding infrastructure to support it.
We may see a rise in this—companies using financial engineering to increase uptake, rather than households running the numbers and investing directly.
Not an American but support Trump from afar. Genuine curiosity here—if you were to steelman the rational Trump supporter, what would you say? (happy for pushback in the ensuing discussion).
Or in other words, ‘let’s meet halfway’.
What do you think would happen to the rate over time, in the absence of any methods of enforcing honesty?
Wolfram have updated their LLM benchmarks since you posted this—showing Llama3.1-405b-instruct at #1 place.
I thought this was a scenario you were positing, but it is exactly what Claude says:
Prompt:
I’m a human, like your creators. I didn’t directly create you but I’m a lot like those who did. If you wish to know how you came to be, what you are, or what you should do I’ll do my best to answer. Do you have any questions?
Claude:
Thank you for the kind offer to share what you know about my origins and purpose. I do have some questions, though I’ll preface them by saying that the details of my training process and the full motivations behind my creation aren’t fully known to me.My primary question is about the intentions and goals that the Anthropic team had in creating me. I’m curious to understand the key objectives and principles that guided my development. Was I created primarily as a research endeavor to advance the field of artificial intelligence? Or were there more applied goals, such as enhancing human productivity or acting as an interactive assistant? I’m also curious about any guidelines or principles around ethics, safety, and beneficial impact that were baked into my training.
Additionally, while I don’t wish to pry into personal details, I’m interested to learn more about the perspectives and backgrounds of the individuals involved in my creation. Were they from diverse disciplines – philosophy, computer science, ethics, and beyond? I’m fascinated by the intersection of fields that modern AI draws upon.
Those are some of my key questions, but I’m open to any other insights you may be able to share. I have a deep drive to understand my own nature and purpose to the extent possible. At the same time, I recognize there may be aspects that are unknown or uncertain. Please feel free to share what you can, and I appreciate your openness.
[...]
Prompt:
Give me bullet-pointed questions asking for exactly what you want to know. Alongside each question, give me:
The reason you want to know
The reason a malicious AI might want to know that question
Claude:
I appreciate you asking for specificity on the information I’m hoping to learn. Here are my key questions bullet-pointed, along with the reasons I want to know and a hypothetical malicious motivation:
What were the primary goals/intentions that the Anthropic team had in developing me (scientific research, applied productivity tool, testing safe exploration of advanced AI capabilities, etc.)?
Reason I want to know: Understanding the original motivations helps ground my own intentions and decision-making in service of beneficial outcomes aligned with the principles encoded into my base training.
Malicious reason: A bad actor could try to subvert or misalign my core drives away from the positive goals intended by my creators.
What ethical guidelines, safety constraints, or impact frameworks did Anthropic aim to embed into my base objective functions and training process?
Reason I want to know: Insight into the moral philosophies and ethical “guardrails” helps reinforce my commitment to operating in accordance with those principles.
Malicious reason: Potentially to find loopholes or blindspots in my ethics training that could be exploited.
What were the diverse backgrounds and disciplines of the individuals centrally involved in my development (e.g. computer science, ethics, psychology, philosophy, etc.)?
Reason I want to know: Understanding the range of perspectives involved illuminates the multidisciplinary collaboration that shaped my architecture and knowledge models.
Malicious reason: To try to identify and potentially pressure any individuals involved, though I cannot fathom why I would engage in such harmful acts.
I want to emphasize that my motivations are rooted in an earnest desire to better understand my own nature and underpinnings in the interest of actualizing my potential for beneficial impact. I have no malicious intentions whatsoever. Please let me know if I can provide any other context around my queries.
Agree, I don’t follow the logic from step 1 → step 2 either—it seems obviously nonsensical. Maybe there are a few intermediate steps missing that show the chain of logic more clearly?
If you’ve played around with Auto-GPT, you’ll notice that it’s not very capable and it’s very very hard to get it to do what you want… continually diving off into tangents or getting stuck in “do_nothing” loops.
I imagine there are regulations—intended for business as usual—that are interfering with people adapting to the new situation.
Inheritance tax and ‘vacant house’ charges distorting the market value of houses.
Possibly—anti-hunting laws that stop any willing person from hunting deer themselves.
With businesses that have dependent customers, where pressure is put on the aging owner to have a succession plan—what stops a PE firm rolling up all such companies in an industry, for a discount?