typon

Here's a better idea: make public universities better and more competitive.

Kinko's is not the pinnacle of human engineering - TSMC is. A slight difference there.

A significantly worse model that says naughty words.

AI Safety is a science fiction created by large corporations and useful idiots to distract from working on boring, real AI safety concerns like bias, misinformation, deepfakes, etc.

Am I tripping or is the linked manual for the Mark II, not the Mark I?

This says less about consumer habits than about thr quality of EVs being sold in Europe and North America

I think solving ARC-AGI will be necessary but not sufficient. My bet is that the converse will not be true - a model that will be considered "AGI" but does poorly on ARC-AGI. So in that sense, I think this is an important benchmark.

Polar icecaps shuddering at the thought

Anyone know John Carmack's status on his AGI company?

Maybe that proves his point?

That's an even better example. Although I can see Ted Chiang arguing that the really good directors painstakingly redo shots, cut, edit etc. Maybe the essence of art is the grind? Not sure.

"accounting fraud" is so quaint compared to what it takes to outrage people these days

My friend is a very good painter, but not a famous painter. In the art world, its an open secret that most famous painters, especially ones that are old, don't really paint much. They hire "ghost-painters" to do the actual work for them, and they simply set the direction of the art pieces and collaborate with the hired-on-contract ghost-painters. My friend has painted for a bunch of these artists and when I ask her whether its unethical, she just shrugs her shoulders because she needs to pay for rent but also, importantly, she thinks that the painting really does belong to the artist setting the direction - that she's merely doing the grunt work.

Are the thousands of choices of which brush strokes to put where actually the seed of creativity? According to some artists - not really.

This is brilliant

> soon they will be equal

That's a prediction, one that I hope comes true. However, I have to be realistic and given the track record I am not too hopeful.

I would love it if there was a working autonomous vehicle outside of Waymo. Google's monopoly in search engines has stifled innovation in that space for long enough. The problem is that Tesla just hasn't demonstrated they are capable of doing it. My bet is that achieving autonomous driving with only RGB vision + sensors that are not LIDAR is just too hard.

Tesla has yet to demonstrate autonomous driving capability. A human ready to intervene in long tail scenarios (which is the hard part of the problem) does not constitute autonomous capability, no matter what the marketing material says.

* No ability to browse the site without logging in

* Hundreds of spam account followers

* Sponsored content inserted in replies masquerading as real content

* Random bugs with video content constantly

* Twitter blue boosting replies to the top, making conversations effectively pay to win.

* Bot account spam comprising ~50% of the replies to any popular tweet

Despite the above, Twitter is still the best place on the internet to get the latest news and a feel for the zeitgeist. This to me is a testament of the incredible product created by Jack.

Basically ALL the important inventions and scientific discoveries have been made at publicly funded universities.

Why is this news?

As someone who works in a robotics company, our biggest competition is minimum wage Latinos working in warehouses.

Maybe something like CppFront can allow metaprogramming such that the default constructor is deleted by default on all types and becomes opt-in.

That makes me feel better, thank you.

Does anyone else find it wild that you can make such a powerful poison at home so easily? One that can be administered pretty covertly?

I'm not following. How do you propose search is performed by the ASI designed for "AI Research"? (as proposed by the article)

The cost of the LLM isn't the only or even the most important cost that matters. Take the example of automating AI research: evaluating moves effectively means inventing a new architecture or modifying an existing one, launching a training run and evaluating the new model on some suite of benchmarks. The ASI has to do this in a loop, gather feedback and update its priors - what people refer to as "Grad student descent". The cost of running each train-eval iteration during your search is going to be significantly more than generating the code for the next model.

This is a great example of why the age of truly terrible software is going to be ushered in as LLMS get better.

When the cost of complexity of interacting with an API is paid by the LLM, optimizing this particular part of software design (also one of the hardest to get right) will be less fashionable.

Looking forward to the MKBHD review

The Canadian Dream is to get a great education and then move of the US.

You might want to blame the government or this or that but I think as a Canadian I've finally come to reckon with the fact that it's just not in the Canadian ethos to do risky things like make startups. Of course there are exceptions to the rule but they are very very rare. Canadian investors don't want to take big risks and the Americans are just next door waiting to gobble up the talent in search of capital.

Things are going to get much, much worse than this.