SubiculumCode

And what, to your mind, would classify something as learning? I assume that your position is not the hard "only humans/living creatures can learn"

That is a strong statement. I guess you are telling Machine Learning to go fuck itself.

Never said it was. Still, understanding to what extent the ability of Chinese labs to keep up to western models with much less compute needs to be understood.

They breached some TOS, but your first sentence is pure, over the top flim flam

Moreover, reading a copyrighted book and learning from it is not theft.

There are many frames that one can place upon this issue. They do not contradict the other. There are moral framings (stole the internet so go eff yourselves, is one), but so is national security, and so is the doomer recursive self improvement risk, and then there is the framing purely on what this implies for future AI training.

I mainly focus on the last.

It will be hard for a frontier lab to justify spending the compute and data curation needed to advance AI further if that expenditure can be assimilated into your competitor's products within months/weeks. So reality will present labs with three choices:

A. Cease spending massive amounts of money and compute improving those models.

B. make those improved models more difficult to distill from, either through some regulatory regime, or some technical solution, which seems unlikely to me.

C. making the best models available only to select partners and government.

In all these potential outcomes, China, which lacks compute that U.S. labs enjoy, will likely stop seeing massive improvements in their AI models. Improvements to be sure, but right now they are enjoying gains from distillation AND their own model innovations, and these potential outcomes would largely stop one of those sources.

Anthropic in general seems to have better security...but they also had reported an internal AI gained access to outside email services to contact an Anthropic developer

Huh .. strange. I was behind my university's firewall

Thanks

I mean insurance is very specific industry, and does not easily generalize.

Is that statement based on token price? More and more it seems that $/token hides as much as it reveals. Token efficiency, tokenizer differences, etc. I'm not saying that you are wrong, I am just saying it is becoming a bit more difficult making statements like this without a bit more research.

https://thecastlemap.com/

Peer’s Certificate issuer is not recognized.

HTTP Strict Transport Security: false HTTP Public Key Pinning: false

Certificate chain:

-----BEGIN CERTIFICATE----- MIIDbjCCAlagAwIBAgIIGMReeVlgUbwwDQYJKoZIhvcNAQELBQAwaDELMAkGA1UE BhMCVVMxEzARBgNVBAgMCkNhbGlmb3JuaWExFDASBgNVBAcMC1NhbnRhIENsYXJh MRcwFQYDVQQKDA5JbmZvYmxveCwgSW5jLjEVMBMGA1UEAwwMaW5mb2Jsb3guY29t MB4XDTI2MDEwNTIwMTAyMloXDTI3MDMzMTIwMTAyMlowGzEZMBcGA1UEAxMQdGhl Y2FzdGxlbWFwLmNvbTCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAKx1 FdJskXQ/SUOQA+ClFQFTTP7uhSGQ0HUIPHjexi/uEQxlt6cw3+1GkWkbhvBchyip 83Vz/ysqObYV0RL85pAJJRKy3zazXb2UadY3kuCnNReCscH3TUJ+KRY1siuTVHgl fnMP5axUvrR+k/Lx5Pojr2T46gjpd+93gT9I4W3Sd98MRrOo6xCopicJ1cLadYGg 7BVz1zj1nyyPIXKfHSiEBxahFK7wLG+Uz3TqrNIBHcI7Te3BuIzkM7g4XE+DmnSg jFo4Emf4U10VwI2i+YFi4JpoOr4SFF/ruy7tntsFRIdHXaE5YoDvJu5My9VmGicG WwZSJ9dw9nqBQf6UE48CAwEAAaNpMGcwEwYDVR0lBAwwCgYIKwYBBQUHAwEwHwYD VR0jBBgwFoAUcWvREelOR6JJ2t8FGZq4V7LnXFowLwYDVR0RBCgwJoIQdGhlY2Fz dGxlbWFwLmNvbYISKi50aGVjYXN0bGVtYXAuY29tMA0GCSqGSIb3DQEBCwUAA4IB AQB3Of7y+3aeTWwcta9nbMERIZWw28QGr1bwMDrAY97y7FmToUkGCWw3M5SovEi1 M3p92t1puVyM3A+qxQ2dxIx4jgQJZbnhz7k3gqSUJ7cjVwiL3QJXKoge1zBYL9tA DQ8YeBLoseGjI0hgXeB2QrbrIdMpJ4Hp8uHuDNiyzGDmvg8RuMDCpzfUPfvYQFBD c0Vp+9mf2jRTIBhXslJewM3D26ZiEYbcisR2v3+R+jc0AFkqi2wp7tAU+nG+w4A6 Dy6tLN+EgQ1MBGBrFZFoPO7wRCk1nNItxTGlNlLBvTN9by5i7XE5PNQPOxL/JN3i hRZpkDAQCiMnfFgIGw0exlxY -----END CERTIFICATE-----

Risk? Or correctly assumed truth that surveillance is occurring?

Cherry picked examples don't prove that it isn't context dependent.

How much is Anthropic's price due to inference cost or extra margin they can get away with by having the best model?

I haven't been able to get Firefox GPU acceleration to work on Linux in forever.

Red pilled? Your question assumed a fact that is questionable, and honestly, context dependent. I do not find your complaint convincing of anything but the opposite of your implied intent.

Are you sure that this applies?

I read some thinking traces someone posted on X, and yeah, near psychosis from refusing to believe this simple of a solution had not been found already

For all thehubbub, as far as I know, all the math breakthroughs via AI that I've heard about have come from Anthropic and OpenAI, not Chinese models. I could have missed those announcements, but one might think that between close to frontier performance plus cheap tokens, that they'd be leading the way on these things.

Lol, you must have some custom instructions on that beast, lol

Instead of monetary rewards, they should preface it with, "This is AI model is not very accurate on these types of questions" and then see how they fare. In some areas, AI is quite accurate, so their experience may lead to different expectations.

still is a shit article.

Speculation: Anthropic and and OpenAI become more like hedge funds, keeping their models to themselves, using their them and their compute for market/event prediction, internal tech development, and AI self improvement. I already wonder if that is what Google is doing. Why release your best models to help the competition? Their published models are good enough for 99% of consumers, and they can leverage their market dominance in other areas to put that consumer model in front of eyes. Then the decision to improve your internal models would both depend on whether you think they'd have value internally.

Yeah, I'm pretty sure AI is going to go insular within the largest companies. This will only be hastened by the growing national security concerns/awareness.

Gemini had a pop-up on my phone today, asking me if I wanted to "bring Gemini up to speed" by importing conversations from my other AI apps. This tells me that Google is threatened, or data hungry as always, or both. Open source AI, Anthropic, and OpenAI, knocking on doors.

Nice! And also, sorry if my comment came off as mean about the site. I don't think so at all...I also don't understand why your first comment on the post went dead.

There is something very interesting about looking at workspaces. As other commenters mentioned? Does your site delve into non-computer workspaces?

no it didn't

Not one creator's workspace is a messy hoard? This is an idealic, oh crap we better clean up before the photographer arrives, fantasy.

Florida is the typical and deserved target.

Fun game. The concept is good. However:

I couldn't find the go button for a while...finally realized it was part of the keyboard...that could be better

I don't get the colors on the keys, as intuitively it seems like it should be the available letters in the current position, but they don't change based on the letter spot I would be currently typing in.

I wanted to share my result with a friend who loves word games, but there was no share button. Also pasting the link does not come up with a web preview in slack/messengers.

It would be nice to be able to go back and play past days.

In terms of difficulty,hard but achievable. Might think on a difficulty slider, or a climb (first solve 5 letter, if won, 6, then 7?