Ariel Nurieli

helping people demand more from themselves // ex military // failed, failed, almost succeeded, kind of succeeded, I think I’m onto something

[Ariel Nurieli @wannaBfounder]

Military taught me, and I taught my soldiers:

The mind is fucking powerful. Forcefully ignore negative shit, and actively pursue positive shit (even false positives!!)

To be able to successfully navigate tough waters you need to be exceptionally optimistic.

Like irrational.

And most important of all, like crazy important…
A goal. Craft a goal that is visual and real. Something you can write that on paper that is measurable, you can explain to somebody and they’ll understand it (even if they don’t believe it’s achievable). Something that you can wake up to and go to sleep too and read and reread as to remind yourself why you’re doing what you do.

As for me, I have written down on a paper that I will have sold my company by the age of 30 so that I can open up the youth leadership organization of my dreams.

The paper has more details than that, but that’s the gist.

To Increase your chances of success, spend time thinking about this and adding to it. Think exceptionally hard about obstacles you’ll have along the way and make them real challenges (people in your life, lifestyle, bad habits).

Then, weak points where you know you’ll fall and will throw off momentum. (for me it’s even a hit of a joint).

And finally think hard about power ups, aka things, people, environments, hobbies, habits, that you know are within your reach and then when you do them, it significantly increases your chances of staying on task and achieving what you need.

And when you’re done with this, remind yourself every single morning and every single night: this is doable, and this will be uncomfortable.

Friction is irrelevant. It coms and goes. Just read the goal every day and every night. Take a break when you reach your goal, or when it’s changed.

[Ariel Nurieli @wannaBfounder]

These AI companies be switching up on you fast

[Ariel Nurieli @wannaBfounder]

How many more awkward “sorry that was an accident” texts will Siri force me to write

[Ariel Nurieli @wannaBfounder]

Didn't know how to color grade a video, so I asked Claude Midway to create an interactive way for me to learn color grading patterns on the exact snippet of film that I want to produce.
Good use case.

Have claude help you make a decision, instead of having it make it for you.

[Ariel Nurieli @wannaBfounder]

The map and territory analogy is perfect.

TLDR the smarter the models, the better they understand what we want and how to get there.

This does not mean that they know exactly what you want.

Its your partner. Work with it to get what you want.

[Thariq is on vacation @trq212]

https://t.co/hPiZr1kG7r

[Ariel Nurieli @wannaBfounder]

19 hard truths on bringing AI into business. Everything I wish I knew before starting my AI consultancy
https://t.co/KKzc4CAekA via @LinkedIn

[Ariel Nurieli @wannaBfounder]

@James_Peters_1 @AndrewYNg No because the randomness is what is the desired output of the application.

[Ariel Nurieli @wannaBfounder]

@VanpeltVentures so cool

[Ariel Nurieli @wannaBfounder]

at all costs, have a clear mind. It is how you will achieve your greatest successes.

https://t.co/Ik4OgfMlTj

[Ariel Nurieli @wannaBfounder]

Trying to make a brief tutorial to explain WTF is AI?

Tried making it mom friendly!

Let me know in the comments what could improve.

https://t.co/iDtQNjSPuB

[Ariel Nurieli @wannaBfounder]

dropping my first open source package for automating testimonials.

Super simple solution for getting you the testimonials you need to grow your business.

https://t.co/UNnpBaF1In

[Ariel Nurieli @wannaBfounder]

@rdominguezibar Beast

[Ariel Nurieli @wannaBfounder]

@MattOg10 Pro tip, just ask Claude to look at all the processes that are dull, and non-relevant.

I would probably use opus for this, as you don’t wanna accidentally kill an important process.

[Ariel Nurieli @wannaBfounder]

think about all the vibe coders who have piled up tons of ghost processes that just draining their battery and slows the computer.

- When you put Claude code in your terminal, you’re actively letting untrained AI control EVERYTHING.

[Ariel Nurieli @wannaBfounder]

Like it’ll sit off a bunch of things, a.k.a., these ghost processes, all to fulfill the task and hand.

But there’s no telling it to kill off, used processes.

Especially when we close the window once we get the result we want.

We closed Claude, not the processes

[Ariel Nurieli @wannaBfounder]

Back baby

[Justin Poehnelt @JPoehnelt]

Two months ago I was fired by Google for creating the Google Workspace CLI. It went viral, hit #1 on Hacker News, gained thousands of GitHub stars and many thousands of actual users in just a couple days.

It was an incredible, confusing journey, from directors and leaders asking what they could learn from the tool to getting grilled by legal about why the Google logo and brand colors are on the Google Workspace GitHub code repositories.

I think the cause was that Workspace and certain leaders (and projects) were afraid of being disrupted. But the fear wasn't specific to my CLI, it was a broader fear in what agents meant for Workspace. Either way, the irony of my termination was the announcement at Google Cloud Next two days before I was fired that an official Workspace CLI was coming.

I want this out there because it is easier for me to explain my story and it is an experience I want to fully own. It's also part of my healing.

Nearly 7 years at Google was an incredible opportunity for me and I was fortunate to have wonderful teammates and a manager that fully supported me through these last few months. Thank you.

[Ariel Nurieli @wannaBfounder]

Native APIs for local agents… so that you can directly interact with agents from remote application.

Ie. your website can interact directly with your local agents.

[Ariel Nurieli @wannaBfounder]

The world belongs to high agency people

[Ariel Nurieli @wannaBfounder]

The people winning with AI aren't the best prompt engineers. They're the best leaders.

[Ariel Nurieli @wannaBfounder]

This guy is single handedly carrying the AI community.
Thank you for being you

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.