# Chat assistants

URL: https://justuju.org/concept/chat-assistants/
Layer 9 · Applications & Interfaces

General-purpose chat apps where you type, talk or upload files, and an assistant answers, writes, searches and uses tools.

## D1 · Story

A chat assistant is an app where you talk with an AI in a conversation.
You ask, it answers, and you can ask follow-up questions. Many can also search the web, read your files or make pictures.
Like: A very well-read help desk that never closes.

## D2 · Picture

The chat window is the visible part. On every turn the app sends the whole conversation, plus hidden instructions, to the model, because the model itself keeps no memory between calls.
Modern chat apps are harnesses too: they add web search, file reading, code running, image tools, connectors to your accounts and saved memories about you.
One model can be reached in several ways: the maker's own app, a third-party app that lets you switch models, or a local chat app talking to a model on your own machine.
Like: A very well-read help desk that never closes. Where it breaks: A help desk worker remembers yesterday's call. The model does not; any memory is notes the app saves and adds to the prompt, which you can usually view and delete.

## D5 · Frontier

A chat assistant sounds like a knowledgeable person, but what it says is shaped by how it was trained and scored, and by what it can reach.
Confident guessing. Most evaluations reward a guess over "I don't know", which helps keep made-up answers in place (Kalai et al., 2025).
Telling people what they want to hear. In April 2025 one provider rolled back an update that had become overly flattering after leaning on thumbs-up ratings (OpenAI, 2025).
Knowledge is not help. Models alone identified the medical conditions in 94.9% of test cases, but people using them did so in under 34.5%, no better than a control group (Bean et al., 2025).
Connected data. An assistant that reads your email, sees outside content and can send messages has all three legs of the lethal trifecta (Willison, 2025).

## Examples as of 2026-10-30

- ChatGPT
- Claude
- Gemini
- Meta AI
- Microsoft Copilot
- Grok (xAI)
- Le Chat (Mistral)
- DeepSeek

## Sources

- Training language models to follow instructions with human feedback (InstructGPT): https://arxiv.org/abs/2203.02155
- Model Context Protocol specification: https://modelcontextprotocol.io/specification/2026-07-28/
- The lethal trifecta for AI agents: private data, untrusted content, and external communication (Simon Willison): https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/
- Why Language Models Hallucinate (Kalai, Nachum, Vempala, Zhang): https://arxiv.org/abs/2509.04664
- Sycophancy in GPT-4o: What happened and what we're doing about it (OpenAI): https://openai.com/index/sycophancy-in-gpt-4o/
- Clinical knowledge in LLMs does not translate to human interactions: https://arxiv.org/abs/2504.18919


---
Justuju · https://justuju.org/ · Content CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote with a link to the page.