If hackernews existing during the dawn of personal computing you bet your ass that every single cpu release by intel or ibm or whatever would be front page news and everyone would be talking about how they were going to use computers to automate all paperwork etc etc.

you just need some perspective.

youre projecting a deficiency of the human brain onto computers. computers have advantages that our brains dont (perfect and large memory), theres no reason to think that we should try to recreate how humans do things.

why would you bother with all these summaries if you can just read and remember the code perfectly.

Because the context window of the LLM is limited similar to humans. That’s the entire point of the article. If the LLM has similar limitations to humans than we give it similar work arounds.

Sure you can say that LLMs have unlimited context, but then what are you doing in this thread? The title on this page is saying that context is a bottleneck.

memory is different than context

grok-code-fast-1 is quite nice for this actually, its fast and cheap enough that you dont feel bad throwing entire threads away and trying again.

in cursor you can index docs by just adding a url and then reference it like file context in the editor

yeah once again you need the right context to override what's in the weights. It may not know how to use the responses api, so you need to provide examples in context (or tools to fetch them)

im struggling to understand what the issue with that is

This is just an issue with people who expect AI to solve all of lifes problems before they get out of bed not realising they have no idea how AI works or what it produces and decide "it stops working because it sucks" instead of "it stops working because I don't know what I'm doing"

The issue is people who cant code.

i mean in a way xcode is incredibly valuable product considering anyone who wants to publish to the app store (where they collect insane rent for zero effort) needs to interact with it.

Apple probably views the product as the privilege to develop and publish apps.

I can confirm this. Have a relative that works for Apple. Made the mistake of complaining in front of her about their take on sales (they make six-figures annually from my app). She went off on how much they give us (including Xcode) for the privilege of having an app on their phones.

Still don't have a contact there. Would have thought I would at least get someone there to talk with if issues come up.

so it's the wrong incentive. apple has no need to make it better because people are forced to use it anyways.

the average person doesnt need to do that. The benchmark for "is this response accurate and personable enough" on any basic chat app has been saturated for at least a year at this point.

one of you uses more test-time compute before giving your answer :)

is the first person actually incapable of thinking deeply or do they just prefer the other way/have a lower barrier for what is worth saying.

I find i like to sit back and think in a group conversation and chime in at the right moment with something insightful, where other people are just going next token prediction stream of concious blabbering.

They can think deeply, it just takes active effort, whereas the broad-and-quick approach comes naturally. Sometimes sitting back and listening can be fun for them.

you can try getting into improv comedy to develop this sort of skill. I'm also generally a slow thinker, but i dont actually think we think slower, I think we have too high a barrier for what we allow ourselves to say. We're afraid of making a mistake or saying something stupid, but most people just blurt out the first thought that percolates from their subconscious.

The US govt does that all the time.

are you under the impression this somehow responds to my question?

I’m saying our society has already decided that long before this event.

i was asking a question to a person, not society

isnt the cost of not fixing it > trillions of dollars?