Building AI products
Earth
Joined July 2008
- Tweets744
- Following305
- Followers167
- Likes898
Ratandeep Bansal retweeted
Andrej Karpathy’s 1-hour Stanford lecture on AI engineering is one of the best explanations I’ve seen of how AI systems actually work.
The progression is simple:
10% → LLM
30% → Prompt
50% → Agent
70% → Loop
100% → Graph
The key takeaway:
AI engineering isn’t just about writing better prompts.
It’s about building systems around models — giving them context, memory, tools, feedback loops, and data flows.
“Delete everything, keep Graph.”
Definitely worth watching if you’re building with AI agents.
Watch → Bookmark it
Ratandeep Bansal retweeted
Anthropic senior engineer just released a 1-hour course on building a team of agents with loops & graphs:
• 00:27 - introduction to CLAUDE.md & Plan mode
• 11:24 - building "skills" & "hooks" for Claude agents
• 37:02 - building agents & subagents with Claude
• 52:47 - self-improving loops & graphs for Claude agents
this 1-hour watch will replace a $500 agentic engineering course
watch today, then read how to build a team of self-improving agents that work together
Ratandeep Bansal retweeted
Andrew Ng just released a 2-hour course
On building agentic skills from scratch with Anthropic:
00:00 - How to build agent skills with Claude
22:32 - Claude pre-built skills for AI agents
41:07 - Agentic skills vs tools, MCP, and subagents
01:06:06 - Skills for long-running agents
This 2-hour watch can replace 10 paid courses on building agents
Taught with Anthropic themselves
Bookmark and watch it tonight
Then read the article below
Ratandeep Bansal retweeted
Anthropic and Andrew Ng built an agent that uses 90% fewer tokens from scratch:
they dropped the entire book of Frankenstein into a prompt - 108,000 tokens asked one question
the input dropped from 108,000 tokens to 11
here's how:
step 1 → put everything that never changes at the top - tools, then system, then docs
step 2 → mark where the static part ends. everything above it gets cached
step 3 → one stray space breaks it - and you pay full price again
step 4 → the cache dies in 5 min - every read resets the clock
step 5 → cached tokens don't count against your rate limits. free headroom
most people never touch this - it pays for itself on day one
watch & bookmark - this 1-hour brilliant course ↓
Ratandeep Bansal retweeted
We just launched Monaco. Here’s the exact playbook we used. I hope it's helpful to other founders and startups:
1. The launch video: most launch videos I see orient around some stressed out person at their desk or some philosophical approach to company building. Do not do this.
1/ 📻 I built a 1970s all-night radio station run entirely by ONE ~1B-parameter model on a single GPU.
You tune in. Sam Dusk spins records that never existed, reads YOUR weather tonight, and answers you live when you call. 🧵
3/ Stack: MiniCPM5-1B (GGUF via llama.cpp) on one Modal T4 + Kokoro TTS + faster-whisper.
Every song is synthesized live in your browser from a key/scale/tempo/timbre — no audio files, nothing copyrighted. Tune off-station and AI "ghosts" bleed through the static.
4/ Call in and he'll dedicate the next record to you, by name.
Built for the @huggingface Build Small Hackathon 🌲
🎙️ Tune in: huggingface.co/spaces/build-…
▶ Demo: youtu.be/lA46z2mYjF0
📝 Field Notes: huggingface.co/blog/build-sm…
#BuildSmallHackathon
📻 I built a 1970s all-night radio station run by ONE ~1B-parameter model on a single GPU.
Tune in: Sam Dusk spins records that never existed, reads YOUR weather tonight, and talks back when you call.
🎙️ huggingface.co/spaces/build-…
▶ youtu.be/lA46z2mYjF0
Meet Professor Pip 🔭 a 3D avatar that teaches kids by VOICE! runs in browser, zero GPU animated.
Brain: MiniCPM5-1B finetuned on @modal_labs. Open weights + dataset on @huggingface. Thanks @OpenBMB!
▶ youtu.be/Hh_v5jkdHfw
✦ huggingface.co/spaces/build-…
#BuildSmall #BackyardAI
Ratandeep Bansal retweeted
Memoir (@trymemoir_ai) is an autonomous marketing system for software companies, letting builders focus on building.
It turns your code into demo videos and makes social posts in your own voice, automatically.
Congrats on the launch, @MaanavA51108 & @ZhanJason179!
ycombinator.com/launches/QdO…
Ratandeep Bansal retweeted
Last week we hosted the first ever YC Paper Club in Mountain View.
We brought together great AI researchers and founders to discuss both the state of the art and what it actually takes to get it into production.
Thanks to the following presenters:
0:12 - Intro from YC Visiting Partner @FrancoisChauba1
3:49 - Tanishq Kumar (@tanishqkumar07) — Speculative Speculative Decoding (arxiv.org/abs/2603.03251)
18:33 - Guangyao (Stannis) Zhou (@zhouguangyao) — Diffusion-MPC (arxiv.org/abs/2410.05364)
30:26 - Isaac Ward — LeWorldModeling (arxiv.org/abs/2603.19312)
43:54 - Akshay Vegesna (@akshayvegesna) — Deep Learning is Not So Mysterious or Different (arxiv.org/abs/2503.02113)
51:24 - Konwoo Kim (@konwookim) — Pretraining Under Infinite Compute (arxiv.org/pdf/2509.14786)
Ratandeep Bansal retweeted
how is this even legal. opus is performing significantly worse on hallucinations.
imagine relying on this model for anything mission critical and they can just swap out the model without telling for something so much worse.
genuinely don’t understand how it’s legal.
Ratandeep Bansal retweeted
Few observations from YC W26 Alumni day
1. Founders are optimizing for speed rather than amount -
Everyone wants to close fast snd get back to building and selling.
2. Rounds are either 4 on 40 mill, 3 on 30 mill, 2.5 on 20 mill caps
3. Outliers are getting oversubscribed before the demo day. Good luck getting on their cap tables
4. Hardware and deep tech companies dominate - In the AI world, momentum has moved towards doing harder things passionately rather than building something that looks like an easy win.
5. gstack vs superpowers claud plugin - equal weightage. Superpowers for daily task. gstack for ceo mode.
6. Build energy is ⚡⚡⚡⚡⚡⚡
Ratandeep Bansal retweeted
Best YouTube Channels To Learn AI in 2026 (No BS)
1. Fundamentals – 3Blue1Brown
2. Deep Learning – Andrej Karpathy
3. AI Research – Yannic Kilcher
4. Practical AI – AssemblyAI
5. LLMs – AI Explained
6. ML Theory – StatQuest
7. Papers Simplified – Two Minute Papers
8. GenAI – Matthew Berman
9. AI Agents – Nicholas Renotte
10. Applied ML – Krish Naik
11. PyTorch – Aladdin Persson
12. Math for ML – Serrano Academy
13. Industry Insights – Lex Fridman
14. Real-world AI – DeepLearningAI
Ratandeep Bansal retweeted
Robotics is getting its Raspberry Pi moment.
This solves one of the biggest frictions in Physical AI:
Everything has been fragmented.
Different tools for control. Sensors. Calibration. Learning.
Now it’s one unified stack:
• motor control to VLA models in one system
• built for real hardware, not just simulation
-> norma_core_dev just entered PUBLIC ALPHA.
The real unlock is ElRobot:
• 7+1 DOF arm
• fully 3D printable
• ~220$ cost
• ships with CAD, URDF, full build guide
Not a demo…
Infrastructure.
And the part most people will underestimate:
auto-calibration for any robot
No manual tuning.
No trial and error.
Just:
code → hardware → working system
That removes one of the biggest hidden bottlenecks in robotics.
Which unlocks:
• faster iteration
• cheaper experiments
• real-world data loops
Exactly what Physical AI needs.
We’re moving from:
“Can we make it work?”
to
“How fast can we improve it?”
Thank you so much for sharing, @ErickSky!
📍Repro: github.com/norma-core/norma-…
Credit: normacore.dev
