Hacker Newsnew | past | comments | ask | show | jobs | submit | plufz's commentslogin

I know very little about how the environment where they run these models look, but surely they have access to different tools like vector embeddings with more current data on various topics?

If they could "see" the future and exploit that they'd probably have much higher returns.

I would say that if these models independently could create such high returns all these companies would shut down the external access to the models and just have their own money making machine. :)

56% over 8 months with the constraints provided are pretty good results for Grok.

you can (via the api, or to a lesser degree through the setting in the web client) determine what tools if any a model can use

But isn’t that more which MCP:s you can configure it to use? Do we have any idea which secret sauce stuff they have? Surely it’s not just a raw model that they are executing?

with the exception that it doesn't seem possible to fully disable this for grok 4

which is curiously the best model …

I don’t think you need a conspiracy theory to explain this. This is simply capitalism, a system that seems less and less like the way forward. I’m not against markets, but I believe most countries need more regulations targeted at the biggest companies and richest people. We need stronger welfare states, smaller income gaps and more democracy. But most countries seems to vote in the absolute opposite direction.

The end goal of capitalism is the same as the end goal of monopoly.

1 person has all the money and all the power and everyone else is bankrupt forever and sad.


That sounds like it does prevent cheating? But maybe doesn’t prevent ALL cheats. Or do you mean they work so poorly that it doesn’t make any difference at all?

It makes cheating harder and the timeline to a cheat product gets longer than the iteration speed of anticheat. Kind of like fancy locks don't prevent break ins, just take longer to pick and require more specialised tools.

As they say, locks only stop honest people.

They are wrong, though. Locks also stop people who would happily commit an opportunistic theft but who lack the necessary tools or skills, people who would trespass if they could retain some plausible deniability ("oops, I didn't see the signs" vs. "oops, I didn't realise I wasn't supposed to cut that padlock"), and so on.

The honest people are a larger group than the dishonest people.

And being real, the zero-day cheats are closely guarded and trickled out and sold for high prices as other cheats get found out, so for AAA games, the good cheats are priced out of comfort zone and anyone who attempts the lazy/cheap cheats is banned pretty quickly. A significant portion of the dishonest becomes honest through laziness or self-preservation. Only a select few are truly committed to dishonesty enough to put money and their accounts on the line.

Same way there are fewer murderers and thieves than there are non-murderers and non-thieves (at least in western countries).


I mean it works by someone saying look for DotaCheat4.exe and it searches for it. That’s basically it. Also if your engine has the ability to be hooked into (ahem, gta) it will detect that a process has been attached. It may do some memory scanning if they implemented the allocator from the sdk. What I’m saying is, it’s a crap shoot out there whether the devs did or not. Executives use it as a blanket as to not get sued. “We have anti-cheat”. They can claim it was “circumvented” or whatever. They are all garbage. BattleEye, EasyAntiCheat, Vanguard. If you don’t know, here LL giving a run down.

https://m.youtube.com/watch?v=VtHlMTc8lR4&t=49s


Interesting! In which ways is it better than spotify?

YouTube does a really good job at making playlists on the fly. Choosing the first song on YouTube will get you a playlist of somewhat similar songs. It took me a while to get used to relative to Spotify but now I much prefer the YT method.

Yeah it actually saddens me a little, if using good and correct typography will be avoided because of LLM.


It's already happened unfortunately. LLMs learned to write correctly from people who write correctly. Those people are now being blamed for sounding like AI, when AI actually sounds like them (and probably learned from their work without permission). To avoid they, they write differently.


Em dashes are still appropriate for articles, journals, scientific papers, and other academic or professional writing.

In social media comments they came across as pompous even before LLMs and werent particularly appropriate for casual comments.

Though to be fair some people enjoy coming across as pompous and embrace the 'better than the peasants and their lowly minus sign use' attitude. Makes them feel special or as if their writing is markedly better than those without fancy punctuation. (It isnt).

Also yes, im describing two writers i know that are adamant about the em dash being 'a sign of an intellectual wtiter'...they are insufferable pricks.


I use em dashes for the same reason that I use semicolons: it’s how I’m hearing the sentence in my head as I’m typing it.

I think it’s a bit of a stretch to call a part of grammar pompous. It’d be analogous with me calling your post lazy due to its various typos — I’m not, I just found the comparison apt.


I would love to know what sequence of characters you normally use in place of an em-dash to express the same nuance of relationship and timing.


Not the GP, but I often just use commas and parenthetical asides (like this).

It's a different stylistic choice (em dashes are nice and all), but it's not how I think, and my writing reflects how I think.

I also will often use the fabled semicolon. It's easy to use with contrasting statements, but that's not its only use; I can use them in some situations to elaborate where em dashes are used.

I'm not saying they are a perfect replacement for em dashes (again, em dashes are cool), but it's just always been my personal style.


> I also will often use the fabled semicolon

I regularly use the semicolon, especially in a sentence where this are commas use in another way. In my mind, a semicolon is a "greater" separator than a comma; used to separate parts of thought in the same sentence (vs grouping of items or a pause).


The problem to me doesn’t seem to be the em dashes but rather the multiple people around you that actively talk about “being an intellectual writer“ and how they need to signal it with their choice of punctuation. Frankly they sound ridiculous. But again, that has nothing to do with the actual punctuation itself. Writing off a writing tool because of two people you agree are ridiculous doesn’t seem like the right way to respond to their behavior.

I’ve used it for literally decades for both formal and informal writing. On social media and in text messages. It is a very useful way to communicate/pace your sentences.


But is that really what you are seeing in this HN comment thread? People who seem very well researched in the biochemicals and meta studies of Prozac? I don’t. :)


> Corn, potatoes and wheat are important maybe even oranges, but we could live with a lot less alfalfa and almonds. Both alfalfa and almonds contain a lot of nutrients you dont find in large enough amounts (or at all) in corn and potatoes though. And alfalfa improves the soil but fixating nitrogen. Sure almonds require large amounts of water. Maybe alfalfa does as well? And of course it depends on if they are grown for human consumption or animal.


I have added a ”language-and-tone.md” in my coding agents docs to make them use less unnecessary jargon and filler words. For me this change sounds good, I like my token count low and my agents language short and succinct. I get what you mean, but I think ai text is often overfilled with filler jargon.

Example from my file:

### Mistake: Using industry jargon unnecessarily

*Bad:*

> Leverages containerization technology to facilitate isolated execution environments

*Good:*

> Runs each agent in its own Docker container


I need this badly


Just PM if you want my file as a starting point.


Gist it!



I personally am more afraid of not the exact opposite, but more that AI will cause us to have too little content more than we will have too much. I just notice with myself how much less I’m using Wikipedia or forums, and get my answers from OpenAI or Antrophic. From what I remember it’s a over 20% drop in Wikipedia traffic from monthly 2022 vs 2025.


Primary sources will always exist as the content needs to come from somewhere. Personally, I shy away from ChatGPT answers for the same reason I avoid Wikipedia: you've got to find authoritative information to get reliable answers. I wonder if people who use LLMs don't care about accuracy, or don't know about sourcing?


I don't think that's fully attributable to LLMs.

Attention spans for long-form content at are at all-time lows, judging by the metrics I've seen from various platforms across different media types.


Have needs and feelings? (I mean we can’t KNOW that they don’t and we know of this case of an LLM in experiment that try to avoid being shutdown, but I think the evidence of feeling seems weak so far)


But you can have needs and feelings even without doing thinking. It's separate.


I can imagine needing without thinking (like being hungry), but feelings? How and in what space would that even manifest? Like where would such a sensation like, say, sadness reside?


Emotions tend to manifest as physical sensations, and if you don't think that's true it's likely you haven't been paying attention. See also https://www.theatlantic.com/health/archive/2013/12/mapping-h...


But that is just our nervous system that is located in both the brain and the body, they are obviously one connected system. Sure you can have reflexes and simple learning without a brain, but you need cognition for feelings. That is sort of the definition of what feeling are.

One popular definition: feelings are the subjective, conscious mental experience of an emotion, or the conscious perception of bodily states that arise from physiological and neural responses to stimuli


Do you think animals don't have feelings?


Do you think animals don't think? Because the contention was "you can't have feelings without thinking". I believe it's much easier to convince yourself that animals think than it is to convince yourself that they have feelings (say, it's much easier to see that an ant has a thinking process, than it is to tell if it has feelings).


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

Search: