Recording intelligenceAI-generated brief · check the source for context

Practical Uses of ChatGPT

2:11:12 recording · AUTO · 1 speaker

Watch the original

Executive Summary AI
  • Andre opens a practical follow-up to his Deep Dive into LLMs video, promising examples, settings and the way he actually uses these tools in work and life.
  • He explains the model as a roughly one terabyte file holding about one trillion parameters, built in two stages: a costly pre-training that compresses the internet and leaves a knowledge cutoff, then post-training on human-written conversations that gives it the assistant persona.
  • He urges starting a new chat when the topic changes because context tokens are a precious resource, and being deliberate about tiers: the free tier serves GPT-4.0 mini, $20 per month buys Plus with 80 messages every three hours on GPT-4.0, and $200 per month buys Pro with O1 Pro mode and deep research.
  • Tool use is the next jump: internet search let ChatGPT, Perplexity and Grok answer when White Lotus season three airs while Claude, stuck at its April 2024 cutoff, and Gemini 2.0 Pro could not, and deep research spent about ten minutes and 27 sources writing a report on CA-AKG.
  • He closes with code and modalities: a Python interpreter that does the multiplications the model would otherwise hallucinate, Claude Artifacts building a flashcards app and mermaid diagrams, Cursor's Composer vibe coding a tic-tac-toe game with confetti, and voice input via Super Whisper plus true audio in advanced voice mode.

Brief overview

An LLM is a lossy zip file of the internet; tools and context make it useful.

  1. Start a new chat when the topic changesOld tokens distract the model and cost slightly more to sample, so keep the context window short.
  2. Know exactly which model and tier you are onFree gives GPT-4.0 mini, $20 per month gives GPT-4.0 with limits, $200 per month gives Pro, O1 Pro mode and deep research.
  3. Reach for tools whenever facts are fresh or exactSearch, file upload and the Python interpreter beat the model's hazy recollection — and every citation still needs checking.

Questions this recording answers

10 questions, each answered where it is said
Should I start a new chat when I switch topics?

Yes. A new chat wipes the context window back to zero. Leftover tokens cost you twice: the model can be distracted by them, which lowers accuracy, and every extra token makes sampling the next one slightly slower and more expensive. Treat that window as precious working memory.

What do ChatGPT's free, Plus and Pro tiers actually give you?

The free tier gets GPT-4o mini, a smaller model with fewer parameters that writes less well and hallucinates more. Plus, at $20 per month, gets 80 GPT-4o messages every three hours. Pro, at $200 per month, adds unlimited GPT-4o, o1 Pro mode and deep research.

Why is a chatbot's knowledge out of date?

Pre-training compresses the internet into the parameters and costs many tens of millions of dollars over roughly three months, so it is rerun rarely. GPT-4o was pre-trained many months, maybe a year, earlier, and that date is its knowledge cutoff. Recent facts need tool use instead.

When is a thinking model worth the extra wait?

Thinking models, tuned with reinforcement learning, pay off on difficult math and code, where they raise accuracy, but they can emit tokens for several minutes. For travel advice there is no benefit. Start with the fast non-thinking model and switch when the response looks weaker than it should.

Which chatbots could find the White Lotus season three release date?

ChatGPT recognised the question as recent and searched by itself, and Grok searched too; Perplexity does this by design. Claude answered from its April 2024 last update and has no search tool. Gemini 2.0 Pro Experimental lacks real-time access, while 2.0 Flash answered and showed sources.

What is deep research and what does it cost?

It is internet search plus thinking, rolled out over tens of minutes, listed under ChatGPT's Pro tier at $200 per month. On the CA-AKG question it thought for five minutes across 27 sources and produced a cited report. Perplexity and Grok have versions; treat any report as a first draft.

Why do some chatbots get long multiplications wrong?

Without a tool the model simply predicts the digits. ChatGPT wrote Python and Claude wrote JavaScript, and both were correct. Grok 3 has no interpreter, did it in its head and ended in 060 instead of 120. Gemini 2.0 Pro got one right, then failed a harder one.

How do you read a book or a paper with an LLM?

Put the actual text into the context window. For The Wealth of Nations, written in 1776, the chapter comes from Project Gutenberg, gets pasted into Claude, and is summarised before questions follow while reading. PDFs can be dragged in, though a 30 megabyte paper exceeded Claude's chat length limit.

How can I speak to these apps instead of typing?

The phone app has a microphone icon that transcribes speech into text, and about 80% of queries on the phone go that way. The Mac desktop app has no such icon, so a system-wide dictation app is used, SuperWhisper bound to F5. Product and library names are still typed.

What is vibe coding?

Handing control to Cursor's Composer, an autonomous agent over your codebase, telling it what to do and hoping it works. It runs commands and edits many files at once: asked for confetti on a win, it installed React Confetti, rewrote App.tsx, added CSS and a sound file.

Key Quote

“I am a one terabyte zip file.”

— Andre, Practical Uses of ChatGPT, at 12:22▶ watch at 12:22
Key Quote

“And so think of the tokens in the context window as a precious resource.”

— Andre, Practical Uses of ChatGPT, at 17:35▶ watch at 17:35
Key Quote

“And so again, use it as a first draft.”

— Andre, Practical Uses of ChatGPT, at 50:52▶ watch at 50:52
Key Quote

“And vibe coding just refers to giving control to Composer and just telling it what to do and hoping that it works.”

— Andre, Practical Uses of ChatGPT, at 1:18:17▶ watch at 1:18:17

Cite this

“Practical Uses of ChatGPT.” https://www.youtube.com/watch?v=EWvNQjAaOHw. Transcript summary by WhipScribe, https://whipscribe.com/transcript-pages/practical-uses-of-chatgpt.html.

Report this page
What is wrong with this page?