Hacker Newsnew | past | comments | ask | show | jobs | submit | not-a-llm's commentslogin

all game stores are flooded with AI slop. good luck competing


Indy games markets have been flooded with human slop since long before AI was a thing, the marketing playbook hasn't changed at all, if you just fire games into the void you never had a shot.


it took 1 month to make a human slop, now it takes 1 day


People have been making games in a weekend using asset packs/store resources and engines for a while now. Indie gaming speed ran a lot of the marketing bullshit that every kind of app is suffering through now, and there's already a pipeline in place that works pretty well: itch.io demo -> vlogging/streaming/discord -> steam demo -> early access -> streamer promotion.


them having an actual product with huge demand makes them immune

Just like Tesla was immune from all the naysayers which were saying its a highly unprofitable company which will 100% go bankrupt because its economics dont make any sense, and they lost huge amounts of money shorting the stock


Tesla is profitable, but income per share has been very low relative to share price. People who were short Tesla weren't necessarily betting on a bankruptcy, just that shareholders wouldn't put up with the low ROI for much longer. And depending on exactly when they shorted it, they might've actually made a profit.

In contrast, AI companies that are actually unprofitable are dependent on continuously raising additional money to sustain their operations, so a sudden drop in market confidence could become a self-fulfilling prophecy as it makes it more difficult to raise money which makes the business more risky which decreases market confidence in a downward spiral.

An actual product with huge demand is not enough to avert bankruptcy, you also need to serve that demand profitably (like Tesla does).


they never were

if I sell 1 token at $1 and there are a trillion of them minted, that does not make a $1T market cap

but if you want to prove a point against AI with fake data, sure


If I sell 1 token at $139.14 and 13.17 billion of them minted, that makes a $1.8T market cap.

Space data centres are the panicle of intelligent allocation.


more dense tokens means more stuff to fit into the embedding space which is per token, so more work to disentangle later


you use the wrong word

the Anthropic tokenizer is not worse, its more expensive/verbose


So, worse? Because we benchmark off token use when talking about token use, and everyone else understood that.


The most important feature of a tokenizer is dividing the inputs into independent values that the neural network can work with. It's not the size.


I mean it might lead to better performance on the model side. So the tokenizer is better but more expensive.


It’s better for them


it can be faster, but is it correct?

React was invented because jQuery style state management collapsed into unmanageable code past a certain size


And it wasn't composable, state was kept in the DOM. Even with higher level libs like backbone we didn't have a component-oriented way like now. Each component had to manually manage all of its child components.


It composes to one level. Each <input> is a component. Is that good enough for your application? For many, it is.


Really sorry to have to say this, but I feel like I'm getting underpaid for the work that I do if people make a living clobbering together input elements and calling it web development.


Noeter theorem and energy get complicated in General Relativity, because there is no global time translation symmetry

https://curtjaimungal.substack.com/p/what-is-energy-actually

Ignoring that, saying "energy must exist because of Noether and time-translation" doesn't really tell what energy actually is, not a very satisfactory response


On the contrary, this is best description of what conservation of energy means: all externally generated forces acting on the system do vary with time. Im other words, this is a closed system as defined in thermodynamics.

The nature of energy is really just a result of this property.

The difficulties with energy in GR are related to the fact that those equations alone allow for space-times that are truly bizarre. Some that are technically solutions to the field equations are clearly nonphysical (e.g. Goedel's solutions). There is also a priori no reason why e.g. expanding metrics should actually yield proper conservation of energy.


> Why is this conserved? Because ∇ᵤ(Tᵘᵥξᵥ) = (∇ᵤ Tᵘᵥ)ξᵥ + Tᵘᵥ(∇ᵤ ξᵥ).

Interesting.

> This is an example of a topic that unstructured self study would likely skip.

I'm not studying physics – could you take a look at some unstructured reflexions I wrote a couple weeks ago?

https://news.ycombinator.com/item?id=48699125


bullshitting has never been a crime

if it were, 90% of startups would be illegal and they founders jailable


Fraud is a crime, but admittedly, not all bullshitting is fraud.


unless you signed a contract with your readers, saying "this article was not LLM written" while being written by LLMs is not fraud, just lying which is legal

but we are not even talking about that, most authors do not claim LLMs were not used


Bullshitting has potentially severe consequences to the bullshitter. A good reputation is worth a lot.

Of course, Elon Musk makes many millions with his bullshit. I admit that.


and then some articles will have maybe a small section, 10% of it, LLM generated. And the it really is a superposition of LLM/human written, and flaggable/unflaggable


It's true! Or maybe 20% or 60% or 80%.


talking about data and function shapes is quite common in functional programming world

https://blog.jle.im/entry/functors-to-monads-a-story-of-shap...


Consider applying for YC's Winter 2027 batch! Applications are open till November 2.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: