Jay-F. 😎

If you don’t know what this means, read this: See, most enterprises already live inside AWS. their data is there. their security rules are there. their compliance checks are there. their logging and governance workflows are there. their infrastructure is there. so when a company wants to build on or with AI, the question being asked is not just: “which model is best?” it is: “can we integrate this model without breaking our infrastructure, our security process?” OpenAI frontier models and Codex are now available through Amazon Bedrock essentially means teams can build with OpenAI models inside the AWS environment they already use and trust. less friction. Codex on AWS is also important because coding agents need: access control audit logs secure environments governance compliance reviews deployment rules Basically all the mumbo jumbo extras enterprises use to stay reliable. AWS already has that trust layer. OpenAI has the models. Boom!!! Paradise.

OpenAI frontier models and Codex are now generally available on AWS, giving enterprises a new way to build on Amazon Bedrock with OpenAI through the security, compliance, and governance workflows they already use. This is also the beginning of a broader expansion of OpenAI capabilities on AWS, including future availability for cybersecurity capabilities like Daybreak. openai.com/index/openai-f…

This is what failed nations need. Violent men determined to stop it all at any cost. the rapists. the kidnappers. the government.

@IAmSteveHarvey The little things.

What's a sign that someone was raised right?

@alexabelonix Glad you liked it.

@only1jayf this is the kind of post i like.

RAG: Retrieval-Augmented Generation or retrieval + external knowledge + generation. We’ve all heard of RAG especially if you’re actively using AI and LLMs but do you know how it works? RAG sounds complicated until you explain it like this: an LLM is like a smart kid in an exam. Let’s call the kid Mark. Mark can write well. but he does not always have the textbook. so when you ask him something specific but he doesn’t have the answer, he would guess. RAG gives him the textbook first. this is the flow: 1/ take your documents (Textbook) 2/ break them into small pieces called chunks 3/ number each chunk or rather turn each chunk into numbers called embeddings 4/ store those numbers in a vector database 5/ when you ask a question, turn your question into an embedding too (numbered chunk) 6/ search for chunks with the closest meaning 7/ put those chunks into the LLM’s context (showing Mark the page in the textbook where the answer is located) 8/ asks Mark the LLM to answer the question you asked using them that’s Retrieval-Augmented Generation. retrieval = find the right notes augmented = add the notes to the question generation = write the answer simple example: you ask: “what is our refund policy?” without RAG: Mark the model would guess the answer possibly from general internet knowledge. with RAG: the system searches your actual refund policy, finds the right section, gives it to Mark the model, then Mark the model answers from that text. RAG makes the model look at evidence before speaking. but bad RAG still fails. if the chunks are messy, the search is weak, the documents are outdated, or the wrong text is retrieved, the answer can still be wrong. the LLM (Mark) is the writer or speaker retrieval is the researcher. if the researcher brings bad notes, the writer/speaker will output confident nonsense.

RAG: Retrieval-Augmented Generation or retrieval + external knowledge + generation. We’ve all heard of RAG especially if you’re actively using AI and LLMs but do you know how it works? RAG sounds complicated until you explain it like this: an LLM is like a smart kid in an exam. Let’s call the kid Mark. Mark can write well. but he does not always have the textbook. so when you ask him something specific but he doesn’t have the answer, he would guess. RAG gives him the textbook first. this is the flow: 1/ take your documents (Textbook) 2/ break them into small pieces called chunks 3/ number each chunk or rather turn each chunk into numbers called embeddings 4/ store those numbers in a vector database 5/ when you ask a question, turn your question into an embedding too (numbered chunk) 6/ search for chunks with the closest meaning 7/ put those chunks into the LLM’s context (showing Mark the page in the textbook where the answer is located) 8/ asks Mark the LLM to answer the question you asked using them that’s Retrieval-Augmented Generation. retrieval = find the right notes augmented = add the notes to the question generation = write the answer simple example: you ask: “what is our refund policy?” without RAG: Mark the model would guess the answer possibly from general internet knowledge. with RAG: the system searches your actual refund policy, finds the right section, gives it to Mark the model, then Mark the model answers from that text. RAG makes the model look at evidence before speaking. but bad RAG still fails. if the chunks are messy, the search is weak, the documents are outdated, or the wrong text is retrieved, the answer can still be wrong. the LLM (Mark) is the writer or speaker retrieval is the researcher. if the researcher brings bad notes, the writer/speaker will output confident nonsense.

When I first used ChatGPT in 2022, I got the same feeling I got as a nine year old discovering Grand Theft Auto. It felt ecstatic.

@blueskypinit For a mother whose baby is about to get drowned, she is relatively calm. She should be murderous. It’s a prank but a fake prank???

She was NOT happy 👀

Tested it out. Incredible.

@ZainabBell657 😂. wtf

pro tip: to transfer a file quickly between computers just copy the file you want, disconnect your mouse, plug it into the other pc and paste

If Claude Mythos doesn’t meet the hype, Anthropic will lose market share.

Anthropic pays $750,000+ a year for engineers who can build LLM architectures from scratch. Stanford taught the entire thing in 1 hour lecture & released it for free. Bookmark & watch this today before someone takes it down and read this article below

@BredsguardDalen Grandpa must have been a point near medically insane

Beach hacks to start off the summer

@elonmusk @jamesdouma Violence man. Unnecessary violence.

@jamesdouma What are the most important things we need to fix?

Been making stuff with grok build this last week. The world is feeling very post-scarcity right now.

@tomfgoodwin what agent are you mostly using?

errmmmmm, not to be miserable but has anyone noticed that agentic AI doesn't really work at all. Like the errors compound, fragile integrations ( any external change breaks it ) , observability is an issue, no verification, context loss, the whole thing seems VERY tricky Not sure this can ever be fixed.

The downside of frontier LLMs is that they’re engineered to be too polite.. They lack the human capacity for rudeness. I need a model with Russian energy. That raw truth. Have you seen an ai model insults? Even a five year old has better banters.

Grok voice command recognition could do with a lot more work. Feels slow in the head.

Down to 36% hourly already.

The Codex usage limits have been reset for all paid ChatGPT subscriptions. You should be back to 100% weekly and 100% hourly limits. Let the tokens do incredible things today and have fun.