rishabhaiover

It used to be self-expression in an oddly entertaining way but that Nikita Bier ruined the whole thing with his metrics chasing algorithmic shifts.

I am not seeing it on claude-code yet

AGI maybe not, but it is reaching disruption level intelligence in the SWE domain.

> LLMs alone aren't the way to AGI

Pretraining + RL works, there is no clear evidence that it doesn't scale further.

Calm down, hysteria doesn't serve any side well.

I read this old paper: https://ieeexplore.ieee.org/document/1176257 that compared VLIWs, superscalars and Vector processors for embedded devices and it turns out Vector processors perform the best (for the given benchmark) with the second-lowest power consumption

> "must collect an extra $650 billion in revenue every year" paired with the idea that automating knowledge work can cause a short-term disruption in the economy doesn't seem logical to me.

I find it funny that Microsoft is scaling compute like crazy and their own products like Copilot are being dwarfed by the very models they wish to serve on that compute.

How does scaling compute does not go hand-in-hand with energy generation? To me, scaling one and not the other puts a different set of constraints on overall growth. And the energy industry works at a different pace than these hyperscalars scaling compute.

wasn't in-context learning an emergent behavior a while ago (1-2 years)?

Agreed. These metrics + my personal use convey reliable intelligence over consistent usage. Moving forward, if context windows get bigger and token price lower, I have a hard time figuring out why your argument would be wrong.

I apologize for making that assumption.

What happened to you?

it also has an effort toggle which is default to High

No serious programmer "vibe" codes. I admit creating SaaS may not be feasible with current infrastructure but you can't ignore the insane jump in productivity that these tools can offer with the right scaffolding.

What makes Anthropic seem like early Apple is not just the unique taste, but the courage to stand firm with their vision of what the product should be.

is this the beginning of the end?

maybe eventually, not in the near-term future.

> When one has faith that sustained attention might hold more value than that gained by interruption, they tend to prioritize the former.

I suspect that attention is naturally tuned to work towards genuine interests which may be orthogonal to conventional value producing tasks

It's one thing to be worried about Meta having underlying ownership of a messaging platform like WhatsApp; it's a whole other ballgame to offer that data to Anthropic or OpenAI.

i suspect awareness on supply-chain attacks is already low (though it seems to be increasing in recent times). the attack surface is everything an agent can get their hands on.

That's just sad. Right when I found love in what I do, my work has no value anymore.

God, gone are the days when I’d spend three days writing unit tests and phone it in for the other two just to reach the weekend.

As a student who has used these tools extensively, I can confirm that AI-assistance in learning does more harm than benefit. The struggle to learn, backtracking from an incorrect assumption and reflection after completing the objective are all short-circuited with agentic tool use. I don't have to say that these tools aren't useful, but I wish they wouldn't sell such an utopian dream of productivity. It's good for some, bad for most.

Earlier, I had to only keep my phone away and not open Instagram while studying. Now, even thinking can be partially offloaded to an automated system.

I think there should be an ai-assistance badge on every Github project. I don't want to look at Contribution graph and Commit history and then eventually the source code to find out the same information. What are we hiding now?

A plea to the wise members of this community, please stop using this word. Use degradation or decay but not this.

Unrelated but iOS 26 is so bad and janky that I've finally decided to switch to an android phone. I hate it so much. Thank god I haven't upgraded to Tahoe.

Philip K. Dick and Frank Herbert saw the fundamental invariants of a capitalistic society much more accurately than we can, even with all of our claimed progress.

I don't fully understand the under the hood mechanics of mmap, but I can sense that you're trying to convey that mmap shouldn't be used a blanket optimization technique as there are tradeoffs in terms of page fault overheads (being at the mercy of OS page cache mechanics)

it allows the program to reference memory without having to manage it in the heap space. it would make the program faster in a memory managed language, otherwise it would reduce the memory footprint consumed by the program.

What are some valid sovereign alternatives NHS can use?