rodiger

It's a proxy for generalized reasoning.

The point isn't that LLMs are the best AI architecture for chess.

404

Also the explicit survivorship bias... this would not be near the front page if their predictions were all wrong.

"System structure mirrors organization"

Conway's Law is the general term for this concept https://en.wikipedia.org/wiki/Conway%27s_law

I don't think most consumers are familiar with the iPhone XR. They know iPhone, and maybe iPhone X, but I don't think the naming will be an issue here.

Not true! The ISS has functioned as an outpost for science- 3,000+ experiments and counting

https://www.nasa.gov/missions/station/five-space-station-res...

This doesn't support your assertion- in fact it does the opposite. The definition(s) of inflation has changed over time. That does not make the current definition(s) less correct

Surprisingly, Nintendo has a long history of copyright-striking videos of folks playing their games.

https://www.ign.com/articles/2013/05/16/nintendo-enforces-co...

https://www.merriam-webster.com/dictionary/aesthetic

Definition 1-c:

pleasing in appearance : attractive

Or literally 90% of the YC class in any given year

There's a difference between "doing an illegal thing as a product" and "lying to investors about your product"

And in a hype cycle many many more projects get off the ground that normally, outside of a hype cycle, wouldn't have ever received the requisite funding.

Ideal condition speed vs congestion speed is a different thing- I'm not suggesting they don't face the standard scaling bottlenecks, but it is also true that under similar congestion conditions, the network will run faster as hardware improves.

SQL databases also run slower under high load, which is totally independent of faster hardware allowing for faster databases vs the baseline

Re: independent verification, I agree with you. It's not a great tradeoff. I'm not here to shill Solana as a panacea or anything of the sort, just pointing out clear inaccuracies from the OOP.

The "1-2%" number the author repeatedly cites is misleading at best. These fees do not scale with the amount transacted. BTC tranfers are currently ~$0.80 whether you're moving $1 or $100k, which is a unique feature.

SQL databases will continue to get faster as hardware speeds up - but blockchains only tend to get slower, as the volume of transactions grows.

This is also misleading, as there are systems like Solana which prioritize speed of confirmation, and do indeed get faster as hardware speeds up. It also allows for sub-cent fees (again, not based on a percentage of the amount moved).

"Your margin is my opportunity"

A ton of companies are dumping R&D into designing their own chips with seeming success. Nvidia will be in tough company quite soon, but for now demand greatly outstrips supply.

The token-by-token responses are probably API, while the "instant" loads seem to be cached.

Probably using groq based on speed of response

Cash grab from their VCs is probably more accurate... I have zero doubt some incredible engineering has gone into the product.

However, from what I can tell they were searching for a problem to solve instead of coming with a distinct, compelling, articulable vision of what they wanted to build

At least it'll look nice on a shelf thanks to Teenage Engineering's good work :)

What is a "full" Dyson Swarm?

Nothing with as much pull as CoD though. By the numbers it's not even close

Total speculation, but maybe Apple focus grouped it and saw karaoke has some corny/cringy implicit associations so they wanted some distance

I think it's more of a "China has been known to conduct mass espionage/monitoring of its citizens and abroad. Given that known fact, TikTok poses a concern." Don't think there's hard proof they are doing nefarious things via TikTok.

1. Celsius was a lending platform which allowed customers to earn yield on deposited crypto. It also had exchange features, but the blow-up is primarily due to the lending aspect.

2. The liquidity crisis happened because of off-chain financing which models current lender/borrower agreements.

Looks awesome! Slight naming collision but pretty different industries. https://www.azuki.com/

Falls apart when we get down to binary, but holds across most levels of abstraction

Those are great for larger companies but most smaller places won't have much listed there

You don't need much infra because of how difficulty adjustments work... consistently less hashpower -> it becomes easier to mine

I don't think they're useful to most people right now (or maybe ever), but there are some cool things in DeFi that are normally reserved for large institutions/not possible in traditional finance:

1. Being a liquidity provider to get a cut of exchange fees (https://docs.uniswap.org/protocol/concepts/V3-overview/conce...)

2. Flashloans for arbitrage (https://aave.com/flash-loans/)

3. Autonomous governance for protocol updates (https://compound.finance/governance)

4. Decentralized perpetual contracts (https://www.perp.com/)

5. Decentralized mutual insurance (https://nexusmutual.io/)

Many/all of these depend on speculation to derive value, but there's added value nonetheless

Individual companies/orgs can confer rights to holders through their T&Cs, but that's revocable and hard to prove. For example: https://boredapeyachtclub.com/#/terms BAYC gives full commercial rights to holders.

Would be really interesting to actually see that challenged in a court.

Yeah I agree. Though seems pedantic to imply blockchains aren't useful because they're "just" composed of cryptographic primitives. Shared state under potentially-adversarial conditions isn't an easy thing to achieve, and has usecases for which other P2P networks aren't sufficient.