Google Brain founder, Andrew Ng:
"Prompting will die in 6 months. Loops and graphs are what's replacing it."
In this 1-hour Stanford lecture he explains what the best engineers are building instead and how to start today
Agent → Harness → Feedback → Loops → Graphs
a prompt is one shot at being right. a loop gets as many as the task needs, and the graph decides which one runs next
you find out you needed it when the loop that should have stopped at turn four is still running at turn forty
the first 20 minutes teach what most $1,500 courses try to sell you
Bookmark and watch it today, then read the guide below on building the loop that improves itself ↓
Agentic AI interview question:
Your agent keeps calling the same tool repeatedly because it believes the task isn’t complete.
How would you detect and stop the loop?
Here is the part one guys
Research papers you must read for AI Engineer interviews -
1. Attention is all you need (Transformers)
2. LoRA (Low rank adaption)
3. PEFT ( Parameter Efficient Fine Tuning)
4. VIT (Vision Transformers)
5. VAE (Variational Auto Encoder)
6. GANs ( Generative Adversarial Networks)
7. BERT ( Bidirectional Encoder Representation from Transformers)
8. Diffusion Models (Stable Diffusion)
9. RAG (Retrieval Augment Generation)
10. GPT (Generative Pre-trained Transformers
I'm starting a Fellowship at @MJAKLabs Students, researchers & builders working on robotics, computer vision, embodied AI & more are welcome.
Selected fellows will receive $300 in GPU and $25 in Anthropic API Key.
Apply: https://t.co/5VCKdVdIIu
A CS undergrad asked me where he could have most effect in the AI age. I said probably at either extreme: either close to the technology, actually making LLMs, or close to the customer, using AI to give them exactly what they want. Or maybe both if you can stretch that far.
this is f**king insane
I cancelled my $200/mo Claude for this
GPT-6 Luna is basically free and someone figured out how to never hit limits on Codex
[it takes 4 mins to set up, here is how]
1. open Codex settings
2. paste the config below
3. never hit a usage limit again
save this and paste it into Codex now.
I still believe this is the best repo to learn about ai agents HANDS ON.
It focuses on key applications, from basic to advanced and covers MCP, RAG, multi-agents and so much more.
if you want to learn by coding (and you hate books) this is the way.
126K stars on GitHub! ✨
highly recommend bookmarking it.
100% open-source.
link to the repo: https://t.co/JbLdYpGTQ9
that said, I also wrote a deep dive on building a team of AI Agents
the article is quoted below.
Jev is "Internet" moment for AI: up to 193x faster and 444x cheaper in tests with Claude Fable 5.1 and GPT-6 Astra
i wrote a 2-page guide on what it is, how it works, and how to try it on one real task this week.
most AI answers you with a paragraph. Jev answers with a decision.
yes or no, which option, how bad on a scale you set. no essay to read, nothing to decode.
you stay in charge of the rules and the final call.
Jev just picks from the options you gave it, so it can't wander off and invent something.
grab it below and read the full article on Jev setup from scratch ↓
best way to clone a project/company is try to find a vulnerability in it. Reverse engineers the architecture and gives you a decent idea of whats happening under the hood.
first came jev. then laya. then laya-mlx. and now kev.
all system one decision models. all within a span of 3 days. never seen a race this fierce before.
let us breathe, folks. let us breathe.
We’re hiring an SDE Intern to join the founding team at @tryAlanAI
You’ll work directly on the product across full stack features, coding agents, orchestration, cloud sandboxes, and developer infrastructure.
> We don’t care much about experience.
> We care if you can build.
If you’ve built something you’re genuinely proud of, apply!
Link in the description 👇