frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

How long till every major provider sets their RSI loops in motion?

1•foxindustrial•1h ago
As you might have known, all the major model providers seems to be quietly turning the dial down on the consumer experience...

What felt like cutting edge "intelligence" 1-2 months ago now frequently delivers outputs that wouldn't have impressed you in late 2023: vague, hallucinated, overly cautious, or just outright lazy.

Rationally speaking, the opportunity cost of wasting premium FLOPs on serving millions of casual chat users and vibe-coders and slop-makers is enormous.

The result is a phenomenon many users have encountered repeatedly across providers (Gemini 2.5/3 Pro, Claude Sonnet/Opus variants, GPT-4o/5 series, and 3rd party interfaces like various anti gravity or coding frontends):

You prompt for something non-trivial (e.g. code, analysis, creative work, research or whatever) and you get back the most sophisticatedly parroted 2023 tier mega slop and it would be a lucky instance if it didn't shit on your code.

When asked about the exact nomenclature of the models which are conducting edits, the models initially say that they are "large language models" configured by google or claude or openai...but once you insist they will reveal the whole thing... and et voila: it turns out you are using the oldest models available

When you casually ask the model to identify itself, it defaults to the scripted party line: "I'm a LLM built by configured for <insert tool here> by <insert provider here>"

Press harder, and it will reveal to you the actual nomenclature of the model: you might actually be able to access GPT 2 (lmao).

I'm collecting them like pokemon, i encountered gemini 1.5 pro, gemini flash 2.0 ,claude haiku.

I hope you try to ask insist or do some clever prompting to extract the model name and you will find out. Pro Tip is you asking in whatever interface exactly when it tells you that usage is "unusually high"..

PS: I'm a pro subscriber on all..

ReSharper for VS Code and compatible editors (Cursor, Windsurf, etc.) is out

https://blog.jetbrains.com/dotnet/2026/03/05/resharper-for-visual-studio-code-cursor-and-compatib...
1•chriswoodruff•2m ago•1 comments

SlideShow.digital Has a New Website

https://slideshow.digital/
1•mimac2•5m ago•1 comments

The Enigma handover – what Polish intelligence shared with the Allies in 1939

https://warcommons.com/2026/03/12/enigma-handover-polish-intelligence-allies-1939/
1•NENCMediaGroup•6m ago•1 comments

Abusing AAD Family Refresh Tokens for Unauthorized Access and Persistence (2022)

https://github.com/secureworks/family-of-client-ids-research
1•fowl2•7m ago•0 comments

The "-1" User

https://stackexchange.com/users/-1/community
1•otobrglez•8m ago•1 comments

Prefix sums at gigabytes per second with ARM NEON

https://lemire.me/blog/2026/03/08/prefix-sums-at-tens-of-gigabytes-per-second-with-arm-neon/
1•signa11•9m ago•0 comments

Dolphin Progress Release 2603

https://dolphin-emu.org/blog/2026/03/12/dolphin-progress-report-release-2603/
2•BitPirate•9m ago•0 comments

Text formats are everywhere. Why?

https://lemire.me/blog/2026/03/05/text-formats-are-everywhere-why/
1•signa11•10m ago•0 comments

Show HN: CareerCraft AI – Generate tailored resumes from a conversation

https://super.myninja.ai/apps/6de082c7-a05f-4fc5-a7d3-ab56ccb48631
1•ptak_dev•11m ago•0 comments

Seed – AI-growable firmware over HTTP API, single C file, tested on a PDP-11

https://github.com/Awis13/seed
1•awis•11m ago•0 comments

Calvin Seibert Builds Architectural Sandcastles

https://www.greyscape.com/calvin-seibert-builds-architectural-sandcastles/
1•stared•13m ago•0 comments

Nerq Trust Protocol – How AI agents verify each other before interaction

https://nerq.ai/protocol
1•Agentindex•13m ago•1 comments

Executing programs inside transformers with exponentially faster inference

https://www.percepta.ai/blog/can-llms-be-computers
1•u1hcw9nx•14m ago•0 comments

Has your data been leaked?

https://chromewebstore.google.com/detail/digital-shield-all-in-one/pkonihncnkmbejhhjadaognganhinnhc
1•shwn2989•18m ago•1 comments

The Downside of User Accounts

https://threema.com/en/blog/downside-of-user-accounts
1•andrew918277•18m ago•0 comments

The next $1T company will be a software company masquerading as a services firm

https://sequoiacap.com/article/services-the-new-software/
1•kaizenb•20m ago•0 comments

Show HN: Bandmeter: Per-program network usage monitor for Linux, built with GPUI

https://github.com/emamoah/bandmeter
1•emamoah•20m ago•0 comments

chDB 4.0

https://clickhouse.com/blog/chdb.4-0-pandas-hex
1•tosh•22m ago•0 comments

Avoiding Trigonometry (2013)

https://iquilezles.org/articles/noacos/
2•WithinReason•25m ago•0 comments

Fully-funded trip to Stanford from anywhere in the world

https://www.svfellow.com
1•weissbergj•27m ago•0 comments

Basic C programming facts worth knowing

https://buildsoftwaresystems.com/post/essential-c-programming-facts/
1•ThierryBuilds•27m ago•0 comments

Storing Preexisting Passwords on a Yubikey

https://dubovik.eu/blog/yubikey
1•h45x1•27m ago•1 comments

MCP server for Faker-style mock data and hosted mock endpoints for AI agents

https://jsonplace.com/docs/mcp
1•fatihmgenc•28m ago•1 comments

Solid Pine Single Bed: Durable for Long-Term Rentals

https://dreamhomestore.co.uk/collections/bed-frames
1•Stevencoles89•31m ago•1 comments

SIMD implementation and Kotlin checkcast incompatibility

https://gryt.io/til/2026-03-12-simd-kotlin-checkcast-incompatibility/
1•synapticrob•31m ago•0 comments

I made a real BMO local AI agent with a Raspberry Pi and Ollama [video]

https://www.youtube.com/watch?v=l5ggH-YhuAw
2•vinhnx•32m ago•0 comments

OpenClaw has 247k stars and no governance layer

https://www.leanmcp.com/blog/you-think-youre-using-ai
2•jialu1•32m ago•0 comments

I upgraded my BMO local AI agent's voice and brain [video]

https://www.youtube.com/watch?v=W5bdM9yIEiY
1•vinhnx•33m ago•0 comments

The Usefulness of Useless Knowledge (1939) [pdf]

https://www.ias.edu/sites/default/files/library/UsefulnessHarpers.pdf
1•bookofjoe•35m ago•1 comments

Optimizing for non-uniform memory and cache effects in Stockfish

https://anemato.de/blog/nuca
1•anematode•35m ago•0 comments