Visual Study Guide – keep your notes, but add auto-generated diagrams and images (visualstudyguide.com)

Excited to announce PolyThink Alpha's early access! Our multi-agent AI system fights hallucinations with consensus-driven, accurate answers from multiple models. I'd love for you to join the waitlist at https://www.polyth.ink/ as I'm planning to randomly roll out invites starting May. Feedback will shape our final launch! I'd love thoughts and suggestions too! What would you like to see here?

Comments (10)

stereo · 13d ago

Isn’t this basically the Swiss cheese model? If your two input AIs hallucinate, or your consensus AI misunderstands the input, you will still have confabulations in the output?

TheKelsbee · 13d ago

I have this same thought, and have tried similar approaches.

OP: Have you trained or fine tuned a model that specifically reasons the worker model inputs against the user input? Or is this basically just taking a model and turning the temperature down to near 0?

kuberwastaken · 12d ago

Low temperature, heavy prompting to answer in a structured way. Sadly can't fine train models since this is API based but the approach does work!

kuberwastaken · 12d ago

From all my testing, this never really happened even once honestly, plus the judge model (that I've kept strictly a reasoning model) also evaluates individually before "judging" the consensus.

sks38317 · 13d ago

I’m genuinely interested in how you arrived at the concept of using AI as a method to treat hallucinations. What inspired that approach?

tough · 12d ago

not op but LLM as Judge is a thing https://arxiv.org/abs/2411.15594

kuberwastaken · 12d ago

Honestly, personal use cases. I am a STEM student and deal with a lot of "hard" questions that are about 60% of the time miscalculated by LLMs, I used to manually paste in approaches from say ChatGPT to DeepSeek and now grok and asked them what do you think is better. I created this out of necessity to automate this then realized how cool it can be if it scales further haha

consumer451 · 13d ago

Very interesting. Will this be available as a meta model via API, allowing use in the coding tool of my choice?

kuberwastaken · 12d ago

Eventually yes, that's the plan! It's extremely good with code too, especially with more vague requests, tends to take about 2-3 rounds but almost always gets a great approach.

Meta, App Makers Launch Washington Lobby to Fight Apple and Google (bloomberg.com)

The Mira Pro Color is Boox's first color E Ink monitor (theverge.com)

Brazil to offer tax breaks to lure data center investments, sources say (reuters.com)

Chai by Langbase: Prompt to Agent (chai.new)

No-engine gamedev using Odin and Raylib (zylinski.se)

Anthony Downs on Personalities Within Organizations (arnoldkling.substack.com)

GSAP is now 100% free for all users (gsap.com)

NotebookLM Audio Overviews are now available in over 50 languages (blog.google)

Worries About AI Are Usually Complements Not Substitutes (thezvi.substack.com)

Weekly Scroll: YouTube's AI Problem (infinitescroll.us)

Bio-based adsorbent from modified sphagnum moss for oil-water separation (nature.com)

Why Your Workflows Should Be Postgres Rows (dbos.dev)

Microsoft: Windows Server hotpatching to require subscription (bleepingcomputer.com)

Lessons Learned Generating 5k AI Personas (saxifrage.xyz)

Applying Team Topologies to Marketing and Community (mbbroberg.fun)

The Star Wars: Tales of the Underworld TV show will premiere in Fortnite game (techradar.com)

Visual Study Guide – keep your notes, but add auto-generated diagrams and images (visualstudyguide.com)

Low Background Steel – Content from Before AI (lowbackgroundsteel.ai)

Www.concurrencycontrol.com (concurrencycontrol.com)

Finland restricts use of mobile phones during school day (theguardian.com)

AI Isn't Only a Tool–It's a Whole New Storytelling Medium (every.to)

College Is Obsolete: AI Is Making Apprenticeships Cool Again (betterschooling.in)

Problem with React Update Model (blog.bloomca.me)

Can AI debug problem scenarios in the OpenTelemetry demo application? (relvy.ai)

What should Engineering Managers be doing, anyway? (taoem.com)

Copilot Arena (github.com)

AI Chatbot Leaderboard (lmarena.ai)

Show HN: Board Buddy – open-source board game counter (boardbuddyapp.vercel.app)

The Only Auto Show That Matters: 2025 Shanghai Auto Show [video] (youtube.com)

Limestone University in South Carolina to Shut After Failing to Raise $6M (bloomberg.com)

Free Software Foundation completes its board member review (fsf.org)

Show HN: CoffeeZip – A coffee passport to discover and collect coffee shops (coffeezip.xyz)

Does UK's Online Safety Act cover misinformation? Well, that depends (theregister.com)

Show HN: Native Immediate-Mode UI Library (github.com)

Testosterone gave me my life back (usefulfictions.substack.com)

Startups Are Building Advanced AI Models Without Data Centers (wired.com)

Show HN: Binaural Toneboard (1ps0.info)

JetBrains releases Mellum, an 'open' AI coding model (techcrunch.com)

The Inevitability and Possible Structures of Supercivilizations (1985) [pdf] (cambridge.org)

GhostHub hit 10K lines – now I'm burning out. What would you do?

Love Letters, Governance, Business, and (Seriously) Ignore Me (davedeek.substack.com)

Retrieve.tools – AI company database organized by industry, function, and task (retrieve.tools)

Superconductivity: VanHove singularity confined to topological semimetal surface (nature.com)

The impact of ecosystem nitrogen enrichment on pollen allergy (thelancet.com)

US arrests two alleged leaders of online extremist 764 group (justice.gov)

Our wounds heal slower than the cuts and scrapes of other primates (newscientist.com)

AI Companions Decoded: Common Sense Media Recommends Safety Standards (commonsensemedia.org)

React Rendering as OCaml Modes (uptointerpretation.com)

Valve adds ARM support to Proton (github.com)

RL for Reasoning in LLMs with One Training Example (arxiv.org)

PolyThink: A Multi-Agent AI System to Eliminate Hallucinations

Comments (10)