Twitter Beğeni Hilesi

A simple way to conceptualize agents - connect to any data source, vector stores - code execution engine - orchestrate with other ML models - LLM agnostic - chat or task based agents

Normies struggle with prompting AI Engineers. Thinking like a programmer is a thing that doesn't come to most people naturally. Even if you use English as a programming language, you may need to understand how machines think 🤷‍♂️

Complex agents develop a plan and then execute it step by step The challenge is that even with 90% accuracy - the error adds up across steps So ideally, you want to aim for 95%+ accuracy which is hard to do

AI models mimic human brain function Re-reading helps improve their cognition and reasoning 😃

The biggest beneficiaries of AI are the early adopter non-experts This demo can quickly up-skill themselves and yield AI as a super power Over time, these AI fanatics will be 10x more productive than their counterparts

The straight shot to super intelligence guys raised $1B from a bunch of VCs. Who cares about burning money when it's "superintelligence" that we are after? 😉 😂

Say what you will about Elon... he is the world's best at these complex build-outs! To the extent super-clusters matter when it comes to super-intelligence, it will always be Advantage - Elon Musk! 🫡

AI Agents Are Slowly Coming Alive Autonomous AI agents will soon perform most tasks. Our AI agents can automate various activities such as document generation, on-demand dashboards, report generation, and technical image generation. Here’s how AI Agents in Abacus AI works:

In some sense, OpenAI has truly led to decentralized AI xAI, SSI and Anthropic are all from ex-OAI people As consumers, we should celebrate 🎉

Undeniable Proof That MainStream Media Is Constantly Gaslighting Folks I don't know who most of these folks are, and they have a random Indian actor (Anil Kapoor) on the cover of TIME 100 in AI 🙄 Does anyone except dentists sub to TIME anyway?

OpenAI is considering $2K / month to access their top models. All kidding aside, this will be a Vision-Pro level disaster. I hope it's a joke

Reflection a new OSS model is claiming some insane numbers We are verifying on Livebench AI It it checks out, this is great for open source AI 👏👏

Folks are saying that AI can replace junior engineers Nope! No it can’t. Not yet Just because AI can build simple apps it doesn’t mean it can replace a skilled human completely

Even if Strawberry or Q* or whatever were AGI or ASI, almost no one is going to pay $2000 for it No consumer anyway... What would they do with this super intelligence?

People are making wild claims about AGI and superintelligence and think it will make them rich. I bet no one has any idea what these terms mean. Not even the people who invented them I am willing to pay $100 to anyone who defines one or both terms in the comments and gets 100

I wish the folks over at Google DeepMind would actually invent a drug that cures Cancer or Diabetes Over the last decade they have been producing multiple AI models that they claim will revolutionize healthcare and produce a zillion drugs But where are these miracle drugs?

Regardless of all the hype RAG based chat bots are still the most common applications of LLMs by far…..

If you have worked on a large, complex code base and landed a few PRs You know your job is not going anywhere for a while LLMs struggle, hallucinate and stumble in those situations

We have so much math, probability, statistics, and LLMs, and we are struggling to call this election. Nate Silver has Trump winning; Alan Lichhtman has Harris. The race is very, very tight, and a lot of it has to do with mass-market media gas-lighting My prediction - Trump

10x Knowledge Work With Matrix Agents. These spreadsheet-style agents can operate on multiple documents at the same time. Perfect for stock analysis, ad creation, contract analysis, and accounting work