frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Ask HN: Are algorithmic feeds fundamentally misaligned with user intent?

1•civichalls•1h ago
I’ve been working on a social product recently, and something keeps bothering me.

Most feeds today are optimized for engagement. That makes sense from a business perspective, but it doesn’t really match how I actually want to use them.

Sometimes I want to go deep into a topic. Sometimes I want to see what people in my city are saying. Sometimes I want to understand opposing viewpoints.

But instead of letting me choose that, the system just keeps reinforcing what I’ve already interacted with. It feels less like exploration and more like being nudged into a loop.

After a while, discovery starts to feel narrow. You see more of the same, even when you don’t want to.

So I’ve been wondering, is the issue the algorithm itself, or the fact that users don’t really have control over it?

What would it look like if people could choose how their feed works instead of having it decided for them?

Curious how others here think about this.

Comments

PaulHoule•1h ago
What does “algorithmic feed” really mean?

A lot of people think chronological feed + boosting is “non-algorithmic” but I think that kind of feed privileges stuff that makes people get angry and boost. The same people would say the “explore” feed on Mastodon which shows the most boosted and liked content is wholesome too. I guess if you think being angry all the time is your moral responsibility you would. The rest of us wouldn’t.

To me a good algorithm would be: if OCR reveals text in an image make it disappear (usually the people are angry) if a video looks like an instagram reel with meaningless motion that causes motion sickness make it disappear. I don’t think everyone would agree!

Snowflake Cortex AI Escapes Sandbox and Executes Malware

https://simonwillison.net/2026/Mar/18/snowflake-cortex-ai/
1•mdp2021•1m ago•0 comments

AI is making CEOs delusional [video]

https://www.youtube.com/watch?v=Q6nem-F8AG8
1•yoyohello13•3m ago•0 comments

Nvidia DGX Station – The GB300 Grace Blackwell 748 GB VRAM Workstation

https://www.nvidia.com/en-us/products/workstations/dgx-station/
2•car•4m ago•0 comments

Remove your ring camera with a claw hammer

https://www.hamiltonnolan.com/p/remove-your-ring-camera-with-a-claw
1•gpvos•5m ago•0 comments

Commission presents proposal for EU Inc

https://ec.europa.eu/commission/presscorner/home/en
1•Tomte•6m ago•0 comments

Outpost Moved its Servers to the E.U

https://outpost.pub/you-outpost-moved-to-the-e-u/
2•CqtGLRGcukpy•7m ago•0 comments

Optimizing Industrial Energy Saving with On-Site Photovoltaics

https://www.mdpi.com/2673-9941/6/2/12
1•PaulHoule•7m ago•0 comments

FBI is buying location data to track US citizens, director confirms

https://techcrunch.com/2026/03/18/fbi-is-buying-location-data-to-track-us-citizens-kash-patel-wyden/
2•jbegley•8m ago•0 comments

Design Driven Network Automation [video]

https://www.youtube.com/watch?v=EKMKJG1g-lU
1•mrmrcoleman•9m ago•0 comments

APL Cultivations

https://xpqz.github.io/cultivations/Intro.html
1•tosh•9m ago•0 comments

The #1 Reason Engineers Fail Technical Interviews (Data from 20K+ Sessions)

https://snubber.ai/blog/state-of-hardware-engineering-interviews-2026
1•Andreas2202•10m ago•0 comments

How to know if your prayer is accepted

https://aboutislam.net/counseling/ask-the-scholar/prayer/know-prayers-accepted/
2•marysminefnuf•13m ago•1 comments

Ask HN: How are you protecting yourself from skill atrophy?

2•xpnsec•14m ago•2 comments

Verígrafo – Temporal Knowledge Graph over Spain's 330K+ official government docs

https://verigrafo.com/
1•munirjojoverge•14m ago•0 comments

The Little Red Dot

https://idiallo.com/blog/little-red-dot
1•foxfired•14m ago•1 comments

Russia's Novosibirsk declares cattle disease emergency as culls spark protests

https://www.reuters.com/business/environment/russias-novosibirsk-region-declares-cattle-disease-e...
1•embedding-shape•15m ago•0 comments

Claude Is Getting Better. What About Us?

https://pydantic.dev/articles/learning-in-the-age-of-ai
1•curiouska•18m ago•0 comments

The more I work with AI (LLMs) the more disillusioned I become

2•iRespectYui•18m ago•0 comments

Autoresearching Apple's "LLM in a Flash" to run Qwen 397B locally

https://twitter.com/danveloper/status/2034353876753592372
2•m-hodges•19m ago•0 comments

VHS Slipcover Maker

https://vhs.texs.org/en/slipcover?bg=000000.t1.43.0.137.178.110.t1.43.0.137.178.110&pl=0&logo=0&c...
3•colinprince•20m ago•0 comments

Viewpoints Research Institute

https://www.vpri.org
1•gone35•23m ago•0 comments

MoveSmart – AI moving cost estimator trained on 50k completed moves

https://www.movesmart.co/
1•msheier•23m ago•0 comments

Show HN: HF-agents, CLI extension to find the best model/quant for your hardware

https://github.com/huggingface/hf-agents
1•clmnt•23m ago•0 comments

Show HN: Timeliner.cc – Find Historical Connections

https://timeliner.cc/
1•schnau_software•24m ago•0 comments

Trey the Triceratops

https://www.joopiter.com/stories/trey-the-triceratops
1•gmays•25m ago•0 comments

TigerFS – A Filesystem Backed by PostgreSQL

https://tigerfs.io
3•hahahacorn•25m ago•0 comments

Horizon Worlds Is Dropping VR Support, Going Flatscreen-Only

https://www.uploadvr.com/meta-horizon-worlds-dropping-vr-support/
2•tosh•26m ago•0 comments

Aliens.gov

https://defensescoop.com/2026/03/18/white-house-registers-new-alien-gov-domains-ufo-uap/
2•warthog•27m ago•1 comments

Igbo-Ora, the scientific enigma of the 'twin capital of the world'

https://english.elpais.com/international/2026-03-12/igbo-ora-the-scientific-enigma-of-the-twin-ca...
2•bookofjoe•28m ago•0 comments

Show HN: Browser-based ECU tuning editor for VAG Simos 18 / DQ250

https://theflashbold.github.io/tune-editor/
1•TheFlashBold•30m ago•0 comments