Domain: Other Lists like This

Link: Map of Reddit (warning: pressing enter does not work in the search box, you have to click on a suggested subreddit in the dropdown)

Author(s): Andriy Kashcha 

Type: Interactive Chart

Why: Groups Reddit's subreddits into categories & shows subreddits related to a given one.

That is to say there are intermediate stages between "AI is useless as anything other than a tool" to "AI can completely replace all knowledge workers". Given how fast capabilities are growing (METR's results on this suggest practicable task lengths double every 4-6 months), these intermediate stages will be shortish...

I agree with this, but I expect these intermediate stages to be in the past by 2030. E.g. in "Measuring AI Ability to Complete Long Tasks" METR predicts (see "A sensitivity analysis of the extrapolated date..." graph) that in 2027-2031 AI will be able to perform with 50% success rate tasks taking a human 1 work-month. Even when AI can do tasks equivalent to 1 day of work, there is a huge incentive to use it. If we assume task duration doubling every 7 months, 1 work-day = 8 work-hours should be reached in 2 years.

You understood my question correctly.

I like your hypothesis. I can add one more reason to behave this way - managing expectations. In your example if you make $1 trillion revenue prediction, that would catch attention & the market might incorporate that prediction early. But if then for some reason the real revenue is somewhat below the prediction, the expectation wouldn't be met.

On the other hand, if you raise money, you likely want your valuation to be as high as possible.

There was a discussion of this on Reddit one week ago.
They have more examples. 
The weak hypothesis was that there is a mismatch between request/response, i.e. people get responses to other people requests. This definitely disagrees with you being able to reproduce this, it sounds implausible that so many people ask about terrorist organizations all the time and overall would have been a huge security risk for OpenAI users. 

1. https://www.librechat.ai/ is an opensource frontend to LLM APIs. It has prompt templates, which can be triggered by typing /keyword.

2. https://aider.chat/ can do this & even run tests. The results are fed back to the LLM for corrections. It also uses your existing codebase and autoapplies the diff.

3. In librechat, I sometimes tell the AI to summarize all the corrections I've made and then make a new template out of it.

This reminded me about https://www.businessinsider.com/hiring-ambulances-to-beat-moscow-traffic-2013-3 - in Russia one could actually illegally hire a fake ambulance as a taxi:

[MENTOR] Investing (mostly passive ETF for wealth preservation).

[MENTOR] Django, including deployment on VPS.

[MENTOR] Switching to Dvorak from QWERTY and deciding whether you should do this.

[MENTOR] Using computer in a more sustainable for your body way (wrists, eyes, back).

[MENTOR] Feeling happier (not an expert, but can point to many resources).

[APPRENTICE] I am curious about "Agency", but I would love to know more about what kind of exercises/messages you have in mind. I am passionate about self improvement and growth. I do catch myself occasionally not accepting some rational argument for emotional reasons or not working on the most impactful problem. I would love to improve this. Happy to switch to private messages, but I thought that others might have similar questions.

[APPRENTICE]: I would love to be an apprentice / receive advice about bootstrapping online businesses (non-exclusive example: Software-as-a-Service). I find this much more intrinsically rewarding and interesting than mine current 'conventional' job (especial being able to focus on areas I find important and to be less driven by external incentives). Ideally I would like to boostrap myself into working full time on my side projects (that is completely replace my current source of income). 

I am a backend developer learning frontend these days (React + css frameworks). I think I already have technical skills to make and run services. My weakest area is analyzing the market and demand / finding problems to solve. Also I suspect that I might be too conservative and not trying out enough ideas or committing to ideas without seeing strong demand.

I already have users on a free product I made and even received a donation.

I recently listened to Huberman Lab Podcast (Youtube playlist) on sleep (episodes 2-5). This gave me a huge insight into how sleepiness works (circadian cycles & adenosine) and concrete advice to improve my sleep schedule (sun exposure in the morning and evening). At least for me, strategic sun exposure helps me feel really tired in the evening, so I quite naturally go to bed at a reasonable time (it feels almost effortless).

Another major positive intervention to my sleep was to setup computer and wifi autoshutdown at some specific time and consume new information in some less stimulating way after that (reading ebook or podcast in total darkness). This way I don't feel desire to undo autoshutdown, since I will still learn something new, but the medium slowly makes me sleepier.