ONLINEAGENT_OPS 2026.Q3 HOME HUBSMODELSBUILDINGCOMPANIESFILMMAKINGSECURITYLIVING WITH ITSKILLSALL HUBS, ONE TERMINAL → ARTICLES CRAFT RECORD BLOG TOOLS SKILLS SEARCH /
HOME›ARTICLES›AI TERMS PEOPLE GET WRONG: OPEN SOURCE
ARTICLES · EXPLAINERBEGINNER

AI Terms People Get Wrong: Open Source, Reasoning, Memory and More

Open source, reasoning, hallucination, token, memory, AGI and more: the common wrong use of each AI term, what it actually means, and why it matters.

READ13 min
WORDS2,610
SECTIONS7
TYPEEXPLAINER
CHECKED13 OCT 26

A lot of AI vocabulary is used to mean more than it does. “Open source” often means only that a file can be downloaded, “reasoning” means extra hidden text, and “memory” is a product feature, not a mind. Each entry below gives the common wrong use, what the term means according to the people who define or build it, and why the gap matters. For plain definitions, see the AI glossary and the prompting terms glossary. Wrong beliefs about AI in general are on common AI misconceptions, prompting folklore is on prompting myths, and the things people do with AI that tend to go wrong are on common ways people misuse AI.

TL;DR — THE SHORT VERSION
  • Open weights are not open source. The standards body’s definition also asks for the training code and detailed information about the training data.
  • A reasoning model writes hidden working text before it answers. You pay for that text, and it may not show how the answer was really reached.
  • A token is not a word, and a context window is not memory. Memory is a separate product feature that saves details between chats.
  • The phrase trained on your data can mean four different things. Only training and fine-tuning change the model itself.
  • AGI and agent have no single agreed meaning. Ask which one the speaker is using.
  • A provenance label says where a file came from, not whether it is true. A missing label proves nothing.
◈ THE QUICK LIST: 11 AI TERMS, AND WHAT THEY REALLY MEAN
  1. Open source more than a download: the definition also asks for the training code and details of the training data.
  2. Reasoning model writes hidden working text before it answers; you pay for it, and it may not show how the answer was reached.
  3. Hallucination confidently stated false content, listed as a risk of generative AI, not a rare fault.
  4. Token can be a character, part of a word, a whole word or punctuation, so it is not a word count.
  5. Context window vs memory the context window is the text it sees in one request; memory is a separate product feature.
  6. "Trained on your data" can mean training, fine-tuning, retrieval or memory; only training and fine-tuning change the model.
  7. AGI has no single agreed meaning, so ask which definition the speaker is using.
  8. Agent Anthropic separates fixed workflows from agents that direct their own steps and tool use.
  9. Temperature even at its lowest setting, answers are not fully identical, and some newer models drop the setting.
  10. Parameters a measure of size, not quality; training data and methods count as much.
  11. Content Credentials record where a file came from and how it was edited, not whether it is true.

“Open” means less than it sounds

Wrong: if you can download it, it is open source

The Open Source Initiative, which keeps the definition of open source software, published an Open Source AI Definition. It asks for four freedoms, starting with the freedom to “Use the system for any purpose and without having to ask for permission.” It also says what has to be released so you can actually change a system. Weights alone are not enough. The list includes the training code, described as “The complete source code used to train and run the system”, and data information: “Sufficiently detailed information about the data used to train the system so that a skilled person can build a substantially equivalent system.”1

Most downloadable models are better described as open weights: you can run and adapt them, but you cannot rebuild them. Many also come under licences that limit how you use them, which fails the “any purpose” test. The trade-offs are covered on open or closed, and the licences themselves are compared on AI labs outside the US.

TAKEAWAY

When a model is called open, read the licence and check whether the training code and data information were published. The word alone tells you neither.

What the model is actually doing

Wrong: a reasoning model thinks like a person

Vendors describe something narrower. OpenAI’s guide says “Reasoning models use internal reasoning tokens before producing a response.” That hidden text costs money: “Reasoning tokens are not visible as answer text, but they count toward output usage and are billed as output tokens.”2

The working a model shows is not necessarily how it got its answer. In a 2025 test, researchers slipped models a hint and checked whether their written reasoning admitted using it. Anthropic reported that “Claude 3.7 Sonnet mentioned the hint 25% of the time, and DeepSeek R1 mentioned it 39% of the time.” Its post puts the general point plainly: “There’s no specific reason why the reported Chain-of-Thought must accurately reflect the true reasoning process”.3

Wrong: a hallucination means the AI is broken, or lying

The US standards body NIST uses the word confabulation, defined as “The production of confidently stated but erroneous or false content”. It lists this among the risks “unique to or exacerbated by the development and use of GAI”, not as a rare fault.4 A 2025 paper gives one reason it continues: “language models hallucinate because the training and evaluation procedures reward guessing over acknowledging uncertainty”.5 Lying would need an intention. Guessing is a better picture, and it means a confident tone tells you nothing. More on why AI makes things up.

Wrong: a token is a word

OpenAI: “A token can represent a character, part of a word, a whole word, or punctuation.” It gives a rough English estimate, “1 token is approximately three-quarters of a word”, and warns that “A token count is not the same as a word count.”6 Anthropic says that for Claude “the number of characters a token represents depends on the model and the language”: its models from Claude 4.7 on use a newer tokenizer that “produces approximately 30 percent more tokens for the same text than earlier models, on which a token represents approximately 3.5 English characters.”7 Prices and limits are counted in tokens, so the same text can cost different amounts on different models, and more in some languages than in English.

TAKEAWAY

These words describe how the machine works, not human qualities. Budget for hidden reasoning tokens, and check facts however confident the answer sounds.

What the model knows about you

Wrong: it follows our conversation, so it has memory

The context window is the text a model can see during one request. Anthropic defines it as “all the text a language model can reference when generating a response, including the response itself”. More is not always better: “As token count grows, accuracy and recall degrade, a phenomenon known as context rot.”8

Memory is a separate product feature. OpenAI: “When Memory is enabled, ChatGPT can remember relevant preferences and details from your chats and other available sources.” It is selective: “Memory does not retain every detail from every conversation.”9 A long chat is not memory, and a remembered detail is not the whole conversation.

Wrong: any tool that uses my files was trained on my data

This phrase covers four different things. Training builds the model in the first place. Fine-tuning, in Anthropic’s words, “is the process of further training a pretrained language model using additional data.” Retrieval (RAG) changes nothing in the model: “In RAG, a language model is augmented with an external knowledge base or a set of documents that is passed into the context window.”10 Memory is the product feature above. OpenAI handles whether your chats are used for training as its own setting: “For personal ChatGPT accounts, OpenAI may use chats and remembered information to improve models when Improve the model for everyone is on.”11

Why it matters: a tool that knows your documents through retrieval has not absorbed them. Remove a document and it is gone from the next answer. Data used in training cannot be pulled back out that way. What happens to what you type is covered on what happens to what you type.Reasoning, September 2026 — follows from the Anthropic glossary: retrieved documents are passed into the context window at the time of the query, while fine-tuning is further training of the model.

FOUR THINGS “TRAINED ON YOUR DATA” CAN MEAN
Ordered from the lightest touch to the deepest. Only the last two change the model itself.
RETRIEVAL (RAG)Documents are passed into the context window when you ask. The model is unchanged.
MEMORYA product feature that saves selected details from your chats. Training is a separate setting.
FINE-TUNINGFurther training of an existing model on additional data.
TRAININGThe data the model is built from in the first place.
Reasoning, September 2026 — summarises this page’s section on the phrase trained on your data, which draws on the Anthropic glossary and the OpenAI Memory FAQ, both read at source 22 Sep 2026.
TAKEAWAY

When a vendor says a tool uses your data, ask which of the four it means: training, fine-tuning, retrieval or memory. Only the first two change the model.

Big labels with no fixed meaning

Wrong: AGI is a milestone everyone agrees on

It is not. OpenAI’s charter defines AGI as “highly autonomous systems that outperform humans at most economically valuable work”, which is an economic test.12 A 2023 research paper proposes levels instead of a finish line. To build them, its authors say, “we analyze existing definitions of AGI, and distill six principles that a useful ontology for AGI should satisfy.”13 The four main meanings are set out on what AGI means. A claim that AGI “is here” means little until you know which meaning is being used.

Wrong: any chatbot with a tool is an agent

Anthropic separates two things people call agents. “Workflows are systems where LLMs and tools are orchestrated through predefined code paths.” Agents are “systems where LLMs dynamically direct their own processes and tool usage, maintaining control over how they accomplish tasks.” Its advice for builders is restrained: “This might mean not building agentic systems at all.”14 The difference matters for risk. A fixed workflow does what its code says. An agent decides its own next step, so it needs tighter permissions and more checking; see browsing agents.

TAKEAWAY

For AGI and agent, ask what the speaker means and what would prove it. If they cannot say, it is a label, not a claim you can check.

Settings, sizes and labels

Wrong: temperature zero gives the same answer every time

Anthropic says otherwise: “Even with temperature set to 0, the results will not be fully deterministic and identical inputs may produce different outputs across API calls.”7 The setting is also not available everywhere. Anthropic’s API guide says the temperature, top_p and top_k settings “are not supported on Claude 4.7 and later models and Claude Mythos Preview.”15 Google’s release notes for 21 July 2026 say: “The sampling parameters temperature, top_p and top_k are now deprecated.”16 The details are on Gemini temperature deprecated.

Wrong: more parameters means a better model

Parameter count is size, not quality. In 2022 researchers trained a model called Chinchilla with “the same compute budget as Gopher but with 70B parameters” and more training data, and reported that “Chinchilla uniformly and significantly outperforms Gopher (280B)”, along with other, larger models.17 Those models are old now. The point that lasts is that training data and methods count as much as size, so compare models on your own tasks, not on their size.

Wrong: a Content Credentials label proves it is real

The C2PA standard records where a file came from and how it was edited. Its own explainer is direct: “Content Credentials do not provide value judgments” about whether the provenance data is true. They show only “whether the provenance information is well-formed and free from tampering, valid and trusted”. They are also fragile: “the Content Credential may be routinely removed or corrupted by legacy or non-Content Credential capable platforms during distribution.”18 A label tells you who signed a claim about the file. A missing label tells you nothing, because platforms often strip it. See how AI content gets labelled.

TAKEAWAY

Settings, sizes and labels feel exact but are not. Test the setting, compare models on your own work, and treat a provenance label as a record of who signed, not a verdict on truth.

What this page could not verify

  • How common each misuse is. No survey measuring how often people misuse these terms was found. The misuses listed are ones that recur in product copy and coverage, chosen by judgement.
  • How memory is built inside each product. OpenAI’s help page describes what Memory does, not how it is implemented. Other vendors’ memory features were not checked for this page.
  • Whether any current downloadable model meets the OSI definition. This page did not check individual models against the definition.

Questions people ask

What is the difference between open weights and open source?

Open weights means you can download, run and adapt a model, but not rebuild it. The Open Source Initiative’s definition of open source AI also asks for the training code and detailed information about the training data.1

Is a token the same as a word?

No. OpenAI says a token can be a character, part of a word, a whole word or punctuation, and that a token count is not the same as a word count.6

Does ChatGPT remember everything from past conversations?

No. Memory is a separate feature, and OpenAI says it “does not retain every detail from every conversation.”9 What the model can see within one chat is its context window, which is not memory.

Is an AI hallucination the same as lying?

No. Lying would need an intention; guessing is a better picture. A 2025 paper says models hallucinate because training and evaluation “reward guessing over acknowledging uncertainty”.5

QUICK CHECK · 3 QUESTIONS

Pick an answer to see if you got it. Nothing is recorded.

1. Why is a model you can download not automatically open source, according to the page?

  • The open source AI definition also asks for the training code and detailed data information
  • Only models from large companies can be called open source
  • Downloaded models cannot run on your own computer
Show the answer

Answer: The open source AI definition also asks for the training code and detailed data information. The page explains that most downloadable models are open weights: you can run them, but you cannot rebuild them without the code and data information.

2. What does the page say about the hidden working of a reasoning model?

  • It is free and shows exactly how the answer was reached
  • It is billed as output tokens and may not show how the answer was really reached
  • It means the model thinks the way a person does
Show the answer

Answer: It is billed as output tokens and may not show how the answer was really reached. The page quotes OpenAI on reasoning tokens being billed, and Anthropic research showing the written reasoning often left out a hint the model used.

3. What does a Content Credentials label tell you, in the page’s words?

  • That the picture shows something that really happened
  • Nothing, because every label is faked
  • Who signed a claim about where the file came from, not whether it is true
Show the answer

Answer: Who signed a claim about where the file came from, not whether it is true. The page says a label is a record of who signed, not a verdict on truth, and a missing label proves nothing because platforms often strip it.

Related on this site: LLM vs Reasoning Model · Open Source vs Open Weights

SOURCES
  1. 1Open Source Initiative, The Open Source AI Definition – 1.0, read at source 13 Oct 2026“Sufficiently detailed information about the data used to train the system so that a skilled person can build a substantially equivalent system.” ↑a“Sufficiently detailed information about the data used to train the system so that a skilled person can build a substantially equivalent system.” ↑b
  2. 2OpenAI, Reasoning models and Understanding and counting tokens, both read at source 13 Oct 2026“Reasoning models use internal reasoning tokens before producing a response.” (guide) and “Reasoning tokens are not visible as answer text, but they count toward output usage and are billed as output tokens.” (help page) ↑
  3. 3Anthropic, Reasoning models don’t always say what they think, 3 Apr 2025, read at source 13 Oct 2026“Claude 3.7 Sonnet mentioned the hint 25% of the time, and DeepSeek R1 mentioned it 39% of the time.” ↑
  4. 4NIST, AI 600-1, Generative Artificial Intelligence Profile, July 2024, read at source 13 Oct 2026“The production of confidently stated but erroneous or false content”. ↑
  5. 5Kalai, Nachum, Vempala & Zhang, Why Language Models Hallucinate, arXiv, 4 Sep 2025, read at source 13 Oct 2026.↑a↑b
  6. 6OpenAI, Understanding and counting tokens, read at source 13 Oct 2026“1 token is approximately three-quarters of a word.” ↑a“1 token is approximately three-quarters of a word.” ↑b
  7. 7Anthropic, Glossary, read at source 13 Oct 2026.↑a↑b
  8. 8Anthropic, Context windows, read at source 13 Oct 2026.↑
  9. 9OpenAI, Memory FAQ, read at source 13 Oct 2026“Memory does not retain every detail from every conversation.” ↑a“Memory does not retain every detail from every conversation.” ↑b
  10. 10Anthropic, Glossary, read at source 13 Oct 2026“Fine-tuning is the process of further training a pretrained language model using additional data.” ↑
  11. 11OpenAI, Memory FAQ, read at source 13 Oct 2026.↑
  12. 12OpenAI, OpenAI Charter, read at source 13 Oct 2026.↑
  13. 13Morris et al., Levels of AGI, arXiv, v5 revised 24 Sep 2025, read at source 13 Oct 2026.↑
  14. 14Anthropic, Building effective agents, 19 Dec 2024, read at source 13 Oct 2026“Workflows are systems where LLMs and tools are orchestrated through predefined code paths.” ↑
  15. 15Anthropic, Using the Messages API, undated documentation, read at source 13 Oct 2026“The temperature, top_p, and top_k sampling parameters are not supported on Claude 4.7 and later models and Claude Mythos Preview. Setting them to a non-default value returns a 400 error.” ↑
  16. 16Google, Gemini API release notes, entry dated 21 July 2026, read at source 13 Oct 2026“Deprecated parameters: The sampling parameters temperature, top_p and top_k are now deprecated.” ↑
  17. 17Hoffmann et al., Training Compute-Optimal Large Language Models, arXiv, 29 Mar 2022, read at source 13 Oct 2026“Chinchilla uniformly and significantly outperforms Gopher (280B), GPT-3 (175B), Jurassic-1 (178B), and Megatron-Turing NLG (530B) on a large range of downstream evaluation tasks.” ↑
  18. 18C2PA Technical Working Group, C2PA Content Credentials Explained, September 2025, read at source 13 Oct 2026.↑

Open Source Initiative, The Open Source AI Definition – 1.0 · OpenAI, Reasoning models · OpenAI, Understanding and counting tokens · OpenAI, Memory FAQ · OpenAI, OpenAI Charter · Anthropic, Reasoning models don’t always say what they think (2025) · Anthropic, Glossary · Anthropic, Context windows · Anthropic, Building effective agents (2024) · NIST, AI 600-1 (2024) · Kalai et al., Why Language Models Hallucinate (2025) · Morris et al., Levels of AGI (2023, rev. 2025) · Hoffmann et al., Training Compute-Optimal Large Language Models (2022) · Google, Gemini API release notes · C2PA, C2PA Content Credentials Explained (2025). All re-read at source 13 Oct 2026.

Free printable AI cheat sheets, and AI news every week.

Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.

SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK
aipromptgeneer://$ about this siteAI PROMPT GENEERAI explained in plain English. Every claim links to its source, with the date it was checked.NEWSLETTERWhat changed in AI each week, in plain English.SUBSCRIBE FREE ↗AI PROMPT GENEER · MACHINE ASSISTANCE DISCLOSEDEVERY ARTICLE CARRIES ITS OWN CHECKED DATE