- Видео 23
- Просмотров 11 998
Tommy Codes
США
Добавлен 4 дек 2024
My name is Tommy, and I talk about technology and code. As a software engineer with over 5 years of experience, I create videos on AI, Python, Git, and other technologies that interest me. I enjoy making tutorials that go deep into implementation details, as well as videos that explain complex tech concepts without using code or math. Sometimes, I showcase my own projects or review software. Subscribe if you love tech and want to learn more!
STOP using LLM APIs incorrectly - use this instead!
I talk about the problems that arise when connecting to multiple LLM (large language model) providers directly. I explore LiteLLM, a new piece of open source software designed to solve the problems associated with connecting to multiple AI providers. LiteLLM handles cost tracking, logging and more! I install LiteLLM on my laptop using Docker Compose and give my reaction to the UI and main features of the software.
Просмотров: 5
Видео
Should You Switch to Open Source AI in 2025? Tommy vs Open Source AI
Просмотров 1496 дней назад
Open source artificial intelligence and LLMs are here!I go on a whirlwind tour of open source artificial intelligence models on my macbook air. I do quick demos of the following open source LLMs in Ollama (ollama.com/) / OpenWebUI (openwebui.com/) - qwen 2:0.5b - mistral-7b - deepseek-r:1.5b I give me thoughts and opinions on open source AI as well as show off the various models in action. I al...
3 Things I Wish I Knew Before I Became a Software Engineer
Просмотров 21313 дней назад
I talk about three things I've learned after being a software developer for over 5 years that have really helped me throughout my career. Some of the advice is technical, some of it is more mindset based. If you are a Junior developer you should be able to get something out of this video, regardless of which language you work with: Python, Java, C etc. 00:00 - 00:21 Who am I? 00:21 - 02:36 IDEs...
Navigate in PyCharm like a PRO (JetBrains Navigation Tutorial)
Просмотров 11817 дней назад
I go over how to navigate using the keyboard in PyCharm. I go through the FastAPI source code as an example and go over the most important file and navigation shortcuts that will save you tons of time. All JetBrains IDEs share the main navigation shortcuts, so these shortcuts and tips will work in Intellij, WebStorm etc. APOLOGIES to windows devs, this video only includes the MacOS shortcuts. A...
4 Must Read Books for Junior Developers in 2025
Просмотров 3 тыс.21 день назад
I go over 4 books that I've read that helped me immensely in my career as a software developer. I go over what is included in each book and why you should reach each of the books. If you are still in college, are interning or just started full these books will be extremely helpful for you. Even if you have many years of experience doing software development, these books can still be very helpfu...
How Commits are Stored in git - Become a git PRO
Просмотров 47024 дня назад
Software Engineers or Computer Science students who struggle with git will benefit from this video. I talk about git commit in depth. I explain exactly what a commit object is, where they are stored and how to inspect them. I talk about why git stores its objects in the way that it does in the .git directory. I also give another shoutout to Pro Git, the best book on git ever written. If you are...
What you NEED to know about LLM rate limits
Просмотров 25426 дней назад
Because of the global GPU shortage, all Large Language Model (LLM) inference providers (OpenAI, Anthropic, Google etc) implement strict rate limits. If you aren't careful, these rate limits can destroy your AI application. In this video I talk about why these rate limits exist, how they work in practice for both OpenAI and Antrhopic, and go into some code demos in Python. I explicitly go over t...
Perplexity AI is simpler than you think (explained without code)
Просмотров 7928 дней назад
I go over perplexity.ai, an AI tool that combines LLMs and web search. I explain how this works at a high level without using any code or math. I explain in detail why this AI tool is different than ChatGPT and why it is needed in the first place. Anyone who has used ChatGPT will be able to understand the concepts in this video. Perplexity: perplexity.ai/search/
How AI Learns: explained without any code or math
Просмотров 18329 дней назад
I explain how ChatGPT accumulates knowledge and learns fact at a high level. I do this without using any code or math. I talk about what sorts of questions ChatGPT is good at answering, and which ones it isn't. I explain why AI models like ChatGPT hallucinate and how to know when they are likely to. You will have a better intuition for how to use ChatGPT and other AI models after watching this ...
Abstractions are Lies
Просмотров 14Месяц назад
How Git Branches actually work (easier than you think)
Просмотров 339Месяц назад
I talk about how git branch is actually easy if you understand the core data structure of git. I go over how to create a "manual" branch without using any git commands or APIs. I recommend a book, Pro Git, that is FREE and worth reading cover to cover for any software engineer. Pro Git: git-scm.com/book/en/v2
LLM Pricing Explained (OpenAI API Pricing)
Просмотров 1,4 тыс.Месяц назад
I discuss LLM API pricing using OpenAI as an example. I go over a simple python script that computes the costs of each request and explain what "tokens" are at a high level.
There will NEVER be another programming language
Просмотров 1,1 тыс.Месяц назад
I talk about how AI tools that software developers are addicted to are making it impossible for any new language to ever get adopted. I also discuss why ChatGPT, Claude and other language models are so much better at JavaScript and Python. I finally describe what would need to happen for my claim to be incorrect.
AI is like Dobby from Harry Potter
Просмотров 527Месяц назад
I hate "Prompt Engineering"
Просмотров 272Месяц назад
I go on a rant about why I hate the term "Prompt Engineer". I talk about what prompt engineering actually is and how it is extremely likely that you already do prompt engineering if you've ever used AI. I discuss why I think it's important for people in tech and AI to be more sympathetic to non technical people who might be intimidated by AI.