prolly97

Adding a zero to revenue every year is wild. Regardless of whether you like Anthropic.

>Just like calling everyone commies and nazis I don't think it's helpful to put extreme labels on >everyone<. As it waters out the meaning of extreme labels.

Did they also publish the prompts and models they used for the summaries? It'd be nice to test this on some examples :)

>This use of language waters out the meaning. Just like calling everyone commies and nazis... also I think it'll have the opposite effect you hope.

It's fair that you disagree with AFD, even disagree a lot. But could we please rethink the use of "far", when they have 20% of the seats in parliament?

Their views, are the views of 1/5 people you'll meet in Germany, on average. If we mirrored this logic, the SPD should be "far-left", Grüne should be "far-left pro" and Links should be "ultra far-left pro max".

This use of language waters out the meaning. Just like calling everyone commies and nazis... also I think it'll have the opposite effect you hope.

Some of the smartest people I know are also some of the people that use AI the most. They use their time and energy on solving hard problems. Either in the abstract (research) or more practical problems (entrepreneurship). I completely fail to see how anyone would want to equal researchers and entrepreneurs to criminals.

Hehe, I started out with good and frugal intentions but my $5/month Hetzner VPS has turned into a $100/month bill. I love exploring all kinds of data, but obviously, if you just do that on an ad hoc basis, the data gets stale. So I just setup some continuous workers and nighly batch jobs that ingest and keep certain data up to date :)

I've opened some of the data up to the public: cvrdb.dk for danish companies (working on granting users access to 'full' extracts of certain parts). ratemyaisite.com which is just a tiny subset of a larger project that I've dubbed commoncrawl-explore (cc-explore), which is run on a subdomain of my personal website, but purely for my own sake.

Making your own devex pleasant is the real game changer imho, not the VPS. I've made it easy for myself to deploy new projects using Dokploy (also tried Coolify).

I don't find the maintenance to be a problem - but part of that is not treating hobby projects as something that has to be enterprise-ready :)

I'm currently reading Andrew Carnegies autobiography. Highly recommend it! He lived around the same age.

He benefitted from his employees ideas and initiatives. But that was by design. Both design of the organisation, design of the incentive structures, and the filters used for hiring.

Companies are just a collective of people trying to accomplish something. The idea that an incompetent boss, could attract talent that would be willing to force great outcomes, without due compensation/incentives, doesn't sound like a credible explanation for the mean story of success.

It's definately an article written with emotion - so I think one has to separate underlying arguments from rhetoric. As I read the technological take, it may be framed optimistically, but otherwise directionally correct. Solar power for instance have a very steep adoption curve as we've gotten better at producing solar panels. And batteries allow for async electricity production and consumption, which solves some of the issues that renewables have traditionally had.

And without specifying why the claims are outrageous, you yourself is guilty of not providing any proof.

I fail to see how this moves the debate in a productive direction.

For a meaningful exchange of viewpoints to take place, I find it useful to establish what facts we agree on- and disagree on, and figure out their sources, before we heat up a debate, that is simply downstream of missing or poor evidence. If two opposite parties in a debate, holding strong opinions, but both show up with strong evidence for either side, then both would ideally benefit from being exposed to the other. Closing your eyes to the other sides evidence, while denying them of yours, accomplishes neither your and their goals - if the goal is the get closer to some resemblance of truth.

Making an antithetical statement doesn't invalidate anything by definition. If you have a rebuttal, do as the author, and express it. State specifically what you're opposing, and give your best rebuttal. Then we can consider what set of arguments and data best predict reality.

Thank you! That's actually really interesting reading. Some of the effect sizes are small though - would have been interesting to see a greater variety of benchmarks. Like what's the effect of different incentives given to the human controls (does 20 VS 100 USD change anything? etc.). Anyways, thanks for the links!

Sounds really interesting! Do you have a link to that research? Unless you're an AI, then I'll just accept your statements as fact ;)