conscion

a.k.a "Actions speak louder than words"

> We might not understand particular "emergent" capabilities

The emergent capabilities are the only capabilities we care about

> I'm just wondering why no one tried to RL a model on stuff like "less LOC" and "less overengineering"

Anthropic looked into this and the answer is because it makes the model more stupid

https://www.anthropic.com/research/evaluating-feature-steeri...

All government unions make no sense. You can't have fair labor bargaining when the employees are responsible for electing their own bosses.

Sports betting is the one place that "prediction markets" make the most sense. Instead of there being a house to win against, the market just skims a fee. The line is set organically by betters, not be the house targeting a profit margin, and the house has no incentive to restrict successful betters because they solely profit on flow, not losses.

I think they gloss over a major factor also. They mention:

> Distribution centres are built around motorway junctions (J24 of the M1, the Golden Triangle in the East Midlands) because that’s where road access is.

But they skip _why_ is that road access and motor junction there. It's there because the government decided building roads was something that was it's responsibility. I know this article is UK focused, but for the US if the government decided to build rails also, then they could put them in more convenient places. Instead they allow rail companies to decide which monopoly corridors the companies get to control.

The political solution is 30 year term limits for Senate and House (5 terms and 15 terms respectively). The current system lack of term limits incentivizes inaction

My guess is that they are end-to-end encrypted. And because of Facebook's scale that they're able to probabilisticly guess at what's in the encrypted messages (e.g.a message with X hash has Y probability of containing the word "shoes")

> Actual effective managers do much more

And how many managers are effective vs. only information funnels?

If you're OK with a ridiculously tall phone: https://www.clicks.tech/

These are all commodity use-cases though. Google, Meta, and Anthropic already all have competing products of equivalent quality and customer pricing is being driven down aggressively.

> This is the same as Searles Chinese room. The intelligence isn’t in the clerk but the book. However the thinking is in the paper.

This feels like a misrepresentation of the "Chinese Room" thought experiment. That the "thinking" isn't the clerk nor the book; it's the entire room itself.

> Most of my criticism of Ben’s perspective is against the idea that some kind of emergent morality that we would recognize

I think Anthropic has already provided some evidence that intelligence is tied to morality (and vice versa) [1]. When they tried to steer LLM models morals they saw intelligence degradation also.

[1]: https://www.anthropic.com/research/evaluating-feature-steeri...

It's also the weakening of the dollar. If the dollar is 10% weaker, an international gold seller now needs 10% more dollars to be willing to give you their gold -- which means the "value" in dollar terms is 10% higher.

> The problem wasn't sketchy mortgages, it was the borderline fraudulent financial shenanigans after that.

What do you think was funding the sketchy mortgages? The fraudulent financial shenanigans

> And that's before any discussion about the actual de minimus changes. Changes which will effectively kill the ability for the average American citizen to custom order anything from any other country.

How does it kill the ability to custom order? My understanding of removing de minimus is only that the tariffs now apply to all orders. And because most tariffs now have been set ~30%, an order that was previously $100 is now $130. It seems like many willing to order custom made clothing would also be willing to pay an extra 30%.

It think these "positive news" approaches are always falling prey to stated vs. revealed preference. People's revealed preference is that they want news about _actual_ events, which is why these positive news approaches always stay niche.

People stated reason for not liking news is the stress, attributing this to the negativity of the news. I think a larger issue is the frequency and transience of the updates, leading to oscillations in peoples understand of situations (similar to the car dealership example in the "Thinking in Systems" book).

Modern news networks are always pushing shallow views of new events (i.e. "BREAKING"). Unless someone explicitly follows up on a story, they were only exposed to the crisis and not the resolution of it. I'd love a network that was "yesterdays news" which waited to publish any news until a broader picture of the situation was understood.

> If the number is only 17%, I'm not sure we need to ban them.

Job hunting is a market and the government should tryu to make every market as efficient as possible. Imagine if you went to any other store and 17% of the items you bought were just junk and didn't work.

The robotaxi evaluation only made sense when they claimed to be years ahead of competitors. A robotaxi service operating in a regular taxi market could have large margins. But Waymo is already operating, any Tesla robotaxi company is going to compete away all the margin with Waymo.

What does any of what you wrote have to do with not trusting this new technology?

I don't think they're confused, I think they're approaching it as general AI research due to the uncertainty of how the models might improve in the future.

They even call this out a couple times during the intro:

> This feature was developed primarily as part of our exploratory work on potential AI welfare

> We remain highly uncertain about the potential moral status of Claude and other LLMs, now or in the future

I see this brought up as an issue every time distributing H1-Bs by salary is brought up. What is the issue with big tech "buying them all"?

That Amazon wasn't acquiring it for it's business acumen and was actually acquiring it for some secondary purpose (i.e. market consolidation, data extraction)

The US should really nationalize their rail infrastructure; only the rails themselves, not the railroad companies. The railroads should be treated as transportation infrastructure the same way that the highway system is.

With nationalized rails, the US could set standards for rail-sharing which could help passenger rail and invest in improved rail connections between cities which would improve both freight and passenger rail service.

Currently, each railroad company gets to maintain their own little fiefdoms in certain parts of the US

https://public.railinc.com/about-railinc/blog/who-owns-railr...

So hypocritical of him

"Elon Musk didn't make a billion dollars, his employees did"

This is still possible. There's nothing stopping you from putting comments in the JSON in your file or a webpage. It's just not going to be accepted as valid by JSON parser, which for learning material should be fine.

> The 10th time, the 50th time, I still needed the GPS to guide me.

If anyone else is frustrated by this experience, I've found that changing the setting in Google Maps to have the map always point north has helped me with actually building a mental model of directions. I found instead of just following the line, it forced me to think about whether I'm going north, south, east, or west for each directions.

Hostilities from men towards women are much more dangerous and women have the correct response to be more vigilant about them.

Large corporations buy Windows _because_ they can have this level of control over their machines. The CTOs and auditors want to be able to say they've personally secured their systems using "top of the line" security software.

If anyone is unable to access the otherhand.org site, it's available on the WaybackMachine https://web.archive.org/web/20240604123006/https://www.other...