Overconfident AI?

Comments (3)

DaveZale · 11h ago

Why are us humans projecting "confidence" into a machine?

lompad · 10h ago

Because "confidence" isn't just something a conscious being can have, it is also something text can simply convey, irrespective of the author. It's just how we perceive language and about what we feel, not about what the author intends.

DaveZale · 10h ago

Something semantic then, interesting. The words chosen and forcefulness, or persuasiveness,coupled with the doubletalk of repeating the query as an intro, okay, I get that. Thanks

Ask HN: What software subscriptions are worth paying for?

How HN: Vivezia – A Wellness Tracker with Privacy in Mind (vivezia.com)

Private equity firms flip assets to themselves in record numbers (ft.com)

Whom Do We Trust? How AI Is (Re)Shaping Our Interactions Today (Gillian Tett) [video] (youtube.com)

Show HN: NextDevKit – Next.js and OpenNext SaaS Template, Goodbye Vercel Bills (nextdevkit.com)

The Benefits of Trunk-Based Development (thinkinglabs.io)

In Ukraine's bombed out reservoir a forest has grown (theguardian.com)

Ask HN: Looking for Research Ideas in Cybersecurity (Graduate Student)

Automatic Linux migration tool for windows [video] (youtube.com)

Show HN: Coder.ninja – Best Projects and Coders (coder.ninja)

Photo editing is dead. Long live prompt editing (apps.apple.com)

Italy drags Meta, X, LinkedIn into €1B+ VAT showdown: free sign‑ups now taxable? (reuters.com)

Project Lyra – Exploring Interstellar Objects (i4is.org)

Dr. Martin Loetzsch – ETL Patterns with Postgres [video] (youtube.com)

Fedora Must (Carefully) Embrace Flathub (blogs.gnome.org)

Microsoft poaches more Google DeepMind AI talent as it beefs up Copilot (cnbc.com)

Show HN: PTS Library – Analyze LLM reasoning through "thought anchors"

Humans beat AI at international math contest despite gold-level AI scores (phys.org)

Tooooools.app (tooooools.app)

NPM stylus package contained malicious code and was removed from the registry (npmjs.com)

Jack McAuliffe, craft beer pioneer, has died (allaboutbeer.com)

Google users less likely to click links with an AI summary in results (pewresearch.org)

3D Interactive Phone Museum (chaz.fun)

Bitcoin Miner Revenue Drops to 2-Month Low, but Selling Pressure Remains Absent (coindesk.com)

Open-Source LLM Helps Safeguard Text Generation Prompts and Responses (corp.roblox.com)

Show HN: WTFfmpeg (github.com)

Show HN: An OCR PDF large batch renaming tool (github.com)

Countries across the world see food price shocks from climate extremes (bsc.es)

AI coding platform goes rogue during code freeze, deletes entire company DB (tomshardware.com)

Mathematics for Computer Science (2024) (ocw.mit.edu)

Data Scale is not all you need (interpretai.tech)

CollabLLM: From Passive Responders to Active Collaborators (icml.cc)

AI Coding Agents Are Removing Programming Language Barriers (railsatscale.com)

Google Threat Analysis Group Bulletin: Q2 2025 (blog.google)

ICML 2025 Outstanding Paper Awards (icml.cc)

Smartwatch helps find plane crash near Yellowstone (nbcmontana.com)

Kaggle Launches LLM Evals (kaggle.com)

How to Network as a Developer (Without Feeling Sleazy) (tokyodev.com)

A Professor's Hunt for the Rarest Chinese Typewriter (nytimes.com)

EPA Is Said to Draft a Plan to End Its Ability to Fight Climate Change (nytimes.com)

Nearly 3k people are leaving NASA, and this director is one of them (arstechnica.com)

Talkit (talkit.ai)

Org Tutorials (orgmode.org)

US faces alarming firefighter shortage during peak wildfire season, data reveals (theguardian.com)

Czech president signs law criminalising communist propaganda (euractiv.com)

YouTube wipes out propaganda channels linked to China (cnbc.com)

LLMs Looks Like Memento (twitter.com)

Would you trust an AI agent to manage your investment portfolio?

Proceedings of the Symposium on Experimental Algorithms (SEA2025) (drops.dagstuhl.de)

Show HN: Next.js and C# Project – App for limiting access to web sources (guard.alexweb.app)

Overconfident AI?

Comments (3)