User
Write something
Afternoon Tea is happening in 8 hours
Pinned
๐Ÿ‘‘
โญ
10d โ€ขย 
๐Ÿ“ข Announcements
Welcome to Clief Notes. Here's where to start.
1. Go check out ๐Ÿ“šNavigating The Course to see how to get around and what's here. 2. Start with The Foundation. Concepts, folder architecture, prompting framework. Everything else builds on this. 3. Check in at the bottom of each lesson. Polls, discussion posts, other members working through the same stuff. Use them. 4. When you're ready to build real things join in on our Biweekly competitions and win some real cash. โญ Competitions Mega Thread 5. If you are wanting to dive into the masterminds, grab all the past templates, artifacts and resources. Upgrade and head into the The Vault for Premium and The Drawing Room (VIP) for VIP 6. Post your work. Ask questions. Help others when you can. What are you here to build?
Pinned
๐Ÿ‘‘
โญ
๐Ÿšจ New one in the NLP Logix series is live ๐Ÿšจ
Sat down with Katie Bakewell, a data scientist who's been building this since 2011, back when it was still just called "natural language processing" ๐Ÿงฎ She came up through math (DNA computing, time series on commodities) and thinks about problems like proofs, not recipes. What we get into: ๐Ÿชจ The Indiana Jones "build me a chatbot" boulder she ran from in 2023 ๐Ÿšจ The 7 neural nets that "found" a signal that was completely fake ๐ŸŽ๏ธ A $5M Pagani vs a $100 Toyota, and why "best" is a trap ๐Ÿค– The first chatbot was built in 1966 (ELIZA)... these aren't new ideas ๐Ÿฌ Meta's SAM3 turning hours of labeling dolphin fins into a single prompt ๐Ÿง  Why half the companies asking for AI are solving the wrong problem โ–ถ๏ธ Go watch ๐Ÿ’ฌ Then drop a comment: What surprised you most, or what would you have asked her? Happy learning ๐Ÿ™Œ
Pinned
๐Ÿ‘‘
โญ
๐Ÿค NEW: The Connection Hub is live
๐Ÿ‘‹ Welcome to the Connection Hub - The Vault ยท Clief Notes So I was on the onboarding call this today, and one thing kept coming up that I couldn't stop thinking about: The biggest value of this new age isn't just the tools. It's the people. ๐Ÿ‘ฅ Specifically โ€” people who understand AI the way THIS community teaches it. Not "prompt hacks" and not "10x your output" nonsense, but actually building systems, thinking in workflows, and treating AI like a real part of how you work. That's a rare group. And a lot of you told me the same thing: ๐Ÿ’ฌ "I'd love to work with someone who gets this." ๐Ÿ’ฌ "I want to break into [industry] but don't know anyone in it." ๐Ÿ’ฌ "Who else here does what I do?" So instead of letting those connections happen by accident... I built a place for them. ๐Ÿ‘‡ ๐Ÿ—‚๏ธ๐Ÿ‘‹ Welcome to the Connection Hub - The Vault ยท Clief Notes It's a simple set of pages, split by industry. You find your corner, drop a quick intro about what you actually do and what you're looking for, and connect with people who speak your language.
How do you know which memories your agents actually use โ€” not just which ones you stored?
For anyone running a persistent memory layer that more than one agent reads from โ€” as it grows, how do you know which entries actually get pulled into a run vs. which just sit there adding retrieval noise? Storing is easy and every run tempts you to write more, but a memory that never gets retrieved isn't context โ€” it's surface area the next search has to wade through. I can measure what I wrote; what I haven't cracked is measuring what actually loaded and changed an output. So โ€” do you track retrieval-per-entry and prune what never fires, or is it still by feel? And if you prune, what's your signal that a memory is dead weight and not just rarely-needed?
Recommendations for hardware (Desktop/Laptop)?
So I'm quite conflicted as of now. I've been looking into it for 2 weeks but still can't decide. Then I remembered that I'm part of a 36,000 member community with experts everywhere, so why not ask here. Here's the "context.md": I am a university student but I also love experimenting and working with open source frameworks, LLM's and so on. So a lot of the parts I value is: Portability - A laptop let's me easily go anywhere with all my work, projects, anything and everything. An issue though, because they are laptops, that means it can't handle as much power (watts) coming in = less performance. This again could be a very minor issue, don't have experience with MacBook's but everyone running portable AI friendly local hardware always has a MacBook. - Yet another caveat, it's still possible to have a whole home desktop and connect with it over the internet, but there's always a slight delay and it might not feel that smooth ruining the "flow state" or whatever, not too sure about this as I haven't tested anything other than https://remotly.com for very light usage. Performance - The most important thing is performance of course, I want to run 9B, maybe even Gemma 4 26B on top of tons of locally running systems like my own OS, Hermes, cron jobs and so much more. The overall aspect - Things such as "If I get a laptop, will it sound like an airplane", "Is the plastic good enough to sustain 4-5 years of constant moving in a bag, floors (in a bag) and so on?", "Is the battery life good enough to hold out if I am out the entire day (considering I don't use 100% of the ram of course)", "Cuda would definitely be good, most of the open source LLM stuff is easily supported by Cuda, aka RTX's", "Can 32GB's of DDR5 RAM hold out better than a 16 GB RTX 6 GB VRAM laptop, and is it even worth getting a RTX x on a laptop? Instantly butchering the capabilities?". Now the budget aspect Over the last few weeks I've had come to a few conclusions:
1-30 of 2,428
Clief Notes
skool.com/cliefnotes
What we give away free beats most paid courses. Build durable AI systems with a Marine vet and Edinburgh researcher. 40+ lessons, growing.
Leaderboard (30-day)
1
+1743
2
Bas Rosario
๐Ÿ”ฅ
+1535
3
Curtis Hays
๐Ÿ”ฅ
+1435
Powered by