diamondage

My point is that they will, and that is always going to be useful. Bundles of simplified knowledge (in whatever form) that make research simpler will be useful to share whether humans can understand them or not...

Almost every country in the world has a patent system for a reason...and they are pretty essential for big pharma to function. Granted there should be better calibration of duration - particularly in software, but rewarding first movers is wise in many market conditions. My point regarding "math as a hash list of theorems" is that Theorems will never lose their value, regardless of how powerful AI becomes; they represent compressed knowledge, and as such it will be more efficient for a more advanced AI to query known theorems than reconstruct each one from scratch - in that gap lies a marginal compute saving; an api service or tool call that someone might charge for. Even AI can benefit from specialists...

This misses the raw advantage of a good proof. It makes conceptualization simpler. In some ways math is like a hash list of of theorems. This list makes it simpler to prove other calculations, and will always be useful, to both humans and AI models. I can see two new directions 1 - the creation of specialist theorem models; that can answer questions efficiently about one topic and 2 - we probably need to incentivize and codify ownership of theorems; charging a proportion of the compute saved by using them. Ultimately enabling mathematicians to be paid our true market value!

Perhaps undiscussed in the paper is how the blending of memories may also be better. If evolutionary pressure encourages the structure of brain to optimise for accurate prediction, not memory, then it is quite possible that older individuals may be making better predictions than younger ones, in situations where they don't remember the details...

Its great. The article summaries frequently have odd phrases in them - eg "keep concise", the precise summary" etc. Might want to check that!

In my experience Azure endpoint versus openAI endpoint was way faster and significantly cheaper.

The uptake forced the bigger companies to act. With image diffusion models too - no corporate lawyer would let a big company release a product that allowed the customer to create any image...but when stable diffusion et al started to grow like they did...there was a specific price of not acting...and it was high enough to change boardroom decisions

Why are they wrong? Surely it depends on how you train it?

I'm not sure if you're correct. In fact there has been a revolution in some areas of social science in the last two decades due to the availability of online behavioural data.

The question 'what is science actually for' can be sidestepped. Everyone can agree that it has value, albeit we disagree on the actual value...this is why you need a market. As to how things get priced in such a market, this is a subject for further research...To start, it just needs to tie to something measurable. Heck we've created memecoins with far less backing. Also, we've carved up the conceptual space on a very course grained level with patents, we just need a more immediate, and granular system for doing so...

Actually, the problem is pricing. If we could identify and correctly value new concepts, then we can dispense with citations and just use the correct sum of concept valuations. Perhaps a correctly designed futures market would not only solve getting the right PhD students the right jobs, but bring a lot of speculative capital into fundamental research?

Low latency starlink orders on hold

How well does it work with multiple languages?

Didn't they get their robots from Italy?

Tried creatine?

Is this significant, or does it set a precedent in anyway?

Should have come to the comments first!

I did work on a diagnostic assist tool developed by a large pharma which diagnosed between Asthma and COPD. GPs get this right 52% time, specialists just over 60 and AI came in over 80%...using 12 relatively mundane input variables. I believe there are a lot of these situations, but not clear pathway to FDA approval

yes but you can train relatively dumb AI on high volumes of historical data and presenting factors. This is typically just a 'diagnostic assist', but often significantly better than a human

Underreported competitive advantage of NHS is their unified and comprehensive data

Seems like an odd choice of name from an apparently low attack surface, cybersecurity aware company...

Distilx has a non publically advertised service

Starlink has created a lot of property arbitrage opportunities

Account pooling? Sign up to a pool that distributes your account to others when not being used, in return gives access to multiple streaming platforms on demand...

You can not deliver a pizza based on a 2% chance of fraud. If there's a chance of bomb, shooting etc the same threshold doesn't apply. So the police have completely different criteria.

So Godel proved that there are true theorems that are unprovable. My hunch is that there is a fine grained version of this result <-- that there is a some distribution on the length of the proof for any given conjecture. If true that would mean that we better get used to dealing with long nasty proofs because they are a necessary part of mathematics...perhaps even, in some kind of Kolmogorov complexity-esque fashion, the almost-always bulk of it

would love to see this taken further. I note that structure 101 seems to have been bought by sonar (https://www.youtube.com/watch?v=6Mie3Iya4zE) but they did this back in the day for java etc. would love to have an accessible modern add on with the same concepts of fat, tangle etc.

anywhere in the US; software/messaging apps/security