fp.

Tell HN: Cloudflare silently injects its analytics when you switch nameservers

557•stagas•19h ago•167 comments

Ask HN: What tools are you using for human code review of AI-assisted code?

8•dafelst•20h ago•3 comments

Ask HN: How do you keep up with HN these days?

85•simonebrunozzi•2d ago•142 comments

Ask HN: Where to learn software engineering fundamentals?

10•mstank•1d ago•4 comments

Ask HN: What is your favorite lightweight tool or CLI utility in 2026?

20•pranav_tech26•1d ago•33 comments

Ask HN: When do you think LLM capacity will reach its ceiling?

2•vincenthsin•5h ago•4 comments

Why is the GitHub trending page weirdly excluding DeepSeek projects?

8•tianyicui•15h ago•0 comments

Ask HN: What LLM subscription/provider to use with pi harness?

3•vira28•17h ago•2 comments

Alternative to Internet Archive

19•earworms•2d ago•17 comments

Ask HN: How would you market this service?

4•dmzxnico•1d ago•7 comments

TrueMath – An open-source exact symbolic math tool

3•SAWANs•1d ago•2 comments

Ask HN: Should hard-tech founders join a problem or technology-first PhD lab?

13•misterballer•2d ago•14 comments

Privacy is an underrated advantage in Slack apps

3•alance•1d ago•3 comments

Ask HN: Does a human still review your code?

11•stikit•2d ago•20 comments

old.reddit.com Now Requires a Login

34•chucksmash•9h ago•43 comments

How does an ARP Spoofing Attack work?

6•emn4tor•2d ago•2 comments

Tell HN: In 2021, there were 4.5 jobs per job seeker vs. 2.5 job seekers/job now

11•ed_balls•3d ago•4 comments

Ask HN: What physical publication do you still buy?

8•sergiotapia•2d ago•14 comments

Ask HN: When you sell your company, do you receive the money with a normal wire?

6•supergilbert•1d ago•8 comments

Great music selection without overthinking

4•cr34t1ve•1d ago•9 comments

Live Experiment of the Veblen Effect

5•henryfudgeoffic•2d ago•0 comments

Ask HN: Do you know how much head hunters cost?

5•mojomark•2d ago•12 comments

OpenAI updates its Privacy Policy to include ads

6•1equalsequals1•2d ago•2 comments

New DeepSeek Pricing Published (Peak and Off-Peak)

5•mittermayr•3d ago•5 comments

You've reached the end!

Open in hackernews

Ask HN: When do you think LLM capacity will reach its ceiling?

2•vincenthsin•5h ago

Comments

murzynalbinos•3h ago

IMO we might not hit a hard ceiling on capacity, but we will likely hit a wall with data quality

vincenthsin•3h ago

As human-generated nutritional data runs dry, LLMs will end up eating their own garbage ?

murzynalbinos•1h ago

Yeah, pretty much. Once the good human data runs out and models start training on each other’s output, quality drops fast.

vincenthsin•3h ago

Sorry,I mean capability limits.