I would try to make a platform like Deviantart or Tumblr except OpenAI pays you to make good content that the AI is trained on.

How do you stop people gaming this by feeding it the output of other AIs?

(not to mention defining "good")

You really think an OpenAI-sponsored social network is going to attract people who create and share original content?

Nice in theory but don’t know how practical it is to actually do.

How do you define “good”? Theres obvious examples at the extremes but a chasm of ambiguity between them.

How do you compute value? If an AI takes 200 million images to train, wait let me write that out to get a better sense of the number:

200,000,000

Then what is the value of 1 image to it? Is it worth the 3 hours of human labour time put into creating it? Is it worth 1 hour of human labour time? Even at minimum wage? No, right?

The year of the true Linux phone is near!

Yes, and everywhere else people have to worry about being deported for pointing out Israel's war crimes. At least no one needed to worry about that on 4Chan, but seeing an anonymous racist meme is even worse for people like you.

That is a completely separate problem, and it's dishonest making the comparison. Extremist right wing ideology and genocide is actively advocated on /pol/ as well as anti-Jewish rhetoric. Neo-nazism is not pointing out Israel's war crimes, and pointing out Israel's war crimes is not neo-nazism or anti-Jewish. /pol/ isn't antisemitic for Israel's genocide; they just hate Jewish people.

The Trump administration trying to deport people for doing so is also unjustified. People are freely criticizing Israel on other popular social media (notably TikTok and Instagram) without inciting a modern neo-nazi and right wing movement like what has happened on 4chan in the past 10 years.

I can't access it because of CloudFlare limits. I didn't even know CloudFlare could be a bottleneck or effectively take your site as hostage until you pay.

For me at least, the 10M context window is a big deal and as long as it's decent, I'm going to use it instead. I'm running Scout locally and my chat history can get very long. I'm very frustrated when the context window runs out. I haven't been able to fully test the context length but at least that one isn't fudged.

This will be a bad week in general for Wall Street as the most significant tariffs hit on Wednesday. Because of the chaos or lack of organization from the White House, it will still fall afterwards because markets can't be certain what will happen.

I have a price target for the S&P 500 that I would be very surprised if it stayed lower because the last time the markets were at the price point the economy was worse. All I want to say is there might be another 10% drop before it gets any better.

This is a great predictive piece, written in sci-fi narrative. I think a key part missing in all these predictions is neural architecture search. DeepSeek has shown that simply increasing compute capacity is not the only way to increase performance. AlexNet was also another case. While I do think more processing power is better, we will hit a wall where there is no more training data. I predict that in the near future we will have more processing power to train LLM's than the rate at which we produce data for the LLM. Synthetic data can only get you so far.

I also think that the future will not necessarily be better AI, but more accessible one's. There's an incredible amount of value in designing data centers that are more efficient. Historically, it's a good bet to assume that computing cost per FLOP will reduce as time goes on and this is also a safe bet as it relates to AI.

I think a common misconception with the future of AI is that it will be centralized with only a few companies or organization capable of operating them. Although tech like Apple Intelligence is half baked, we can already envision a future where the AI is running on our phones.

Subject: You're hired! Body: April Fools!

I'm sorry this happens. You can cope by having a few "evangelist" users you communicate with to balance the negativity. There's a bias for comment on code that's not perfect. When everything's perfect, no one has anything to say. Open source bring with it an open quorum.

Github in particular has no way of stopping drive-by comments and issues, right?

You do not have to use Github.

Disable Github issues and use a mailing list instead.

A mailing list doesn't stop drive-by comments and shenanigans. It just raises the barrier to entry.

This isn't going to be a tool recommendation but a learning path one. I recommend choosing a project to work on and then developing it using LLM assisted programming. The more niche and unique the problem, the less likely there's already a solution, and the more likely that the vast training data with a LLM will help you with that niche.

Ideally, you should try out all the tools. Each of them have advantages and disadvantages for each problem within a project. You might find that mixing and matching LLM apps gives you more expertise and skill over just sticking to your favorite.