Pinned Tweet
Excited to share some fun projects I’ve been working on!
If you love dense reading and want a copy, shoot me a DM.
1
157
Khalil Benalioulhaj retweeted
NOW DO OKC
Hawks basketball is coming to Prime Video!
We're the first NBA team to stream local games exclusively on Prime Video, included with your Prime membership at no additional cost for in-market fans! 📺 👀 @NBAonPrime
Khalil Benalioulhaj retweeted
Replying to @googlegemma
is it possible to put this tech inside of Gmail?
Please tell your friends
The LLM delusion is creating a race to the extremely boring middle. Highly skilled humans have turned off their brains in return for the instant gratification of mediocre output. Unskilled humans can generate the same output in seconds. This will lead to huge opportunities.
44
Khalil Benalioulhaj retweeted
The LLM delusion is creating a race to the extremely boring middle. Highly skilled humans have turned off their brains in return for the instant gratification of mediocre output. Unskilled humans can generate the same output in seconds. This will lead to huge opportunities.
Portability matters too much for a feature like this to lure any new meaningful users in.
I think if you already in Codex and are excited for these types applications, you’re either an avid early adopter or not seriously thinking about your operations.
i am putting out my prediction now that this product gets extremely low usage
notion is fine due to their strength of being model agnostic, no serious company is going to move their workflows to a worse version of it's product that is locked in to one provider
57
This brings me so much joy
Again rookie question. All these world class athletes I’ve been watching. How do I continue to watch these guys when they go back to their club teams-Now that I’ve seen them I want to keep watching them-I’m sure this sounds ridiculous but how do you watch these guys on their club teams when you live in the US?
Are all these great players in the same league? When is their season? And how do you buy a package to keep watchin these ballers?
And are the crowds like this for club matches or just the WC?
The energy is what really draws me in (along with the crazy skill).
And lastly how do I find a team to follow?
Other than the obvious guys like Messi/Renaldo-Mbappé the other guy I love is #2 on Morocco Hakimi.
Who is his club team? Maybe that’s my team…he might be my favorite guy I’ve watched the entire tournament.
I’m sorry I’m clueless to all this-just completely uncharted waters for me but I’m hooked and am gonna want more when this is over in a few weeks.
99
One of the best releases from Claude that will drive huge adoption.
Can it run through subscription?
Introducing Claude Tag, a new way for teams to work with Claude.
In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.
55
Introducing Claude Tag, a new way for teams to work with Claude.
In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.
Khalil Benalioulhaj retweeted
You keep texting and calling, screaming like this is all my fault
You’re becoming a challenge I’m calling a timeout like Daigneault
-Drake on ICEMAN 🧊
I'm writing more and more everyday.
It's always been one of the highest leverage activities for me, but even more so with my increased adoption of AI.
I spend way less time clicking buttons now that AI can handle execution.
I write to reflect and to cultivate judgment.
32
Anyone in knowledge work runs into this problem when they try to collaborate with a team of people that’s using Claude Code/Agents.
I’m working on a workspace architecture that has three spaces for the AI Native knowledge team.
Someone will solve this, probably Anthropic. But for now we are solving it for our team with Obsidian + GitHub + Google Drive.
The 3 workspaces:
The Desk
The Shelf
The Table
People will work at their Desk (Obsidian) with Claude Code.
This is the most current and honest place for context.
Teams will reference stable context from the Shelf (GitHub) with Claude Code.
This is where verified and managed context lives: skills, decision logs, project scopes, client details, etc.
When you work on something for a client, your agent grabs context from the Shelf and pairs it with the context at your Desk to be current with your persona context and aligned with the team’s context.
Clients will access, review, and share deliverables at the Table (Google Drive).
Ready to read and share with people they way they are used to accessing deliverables.
It’s no different than how large law firms or consulting firms have worked for decades. They have extensive knowledge bases with shared stable context and best practices so they don’t duplicate work or drift from precedence. They have entire teams dedicated to maintaining this context.
But the small companies never had the resources or bandwidth to maintain a thorough knowledge base. They kept this context in a Google Drive, Notion, Slack, email threads, and in reality, most of it is in their heads.
But with AI, we have agents that can serve as the knowledge management workers small firms could never afford.
Would love to discuss this further with anyone trying to solve this problem or experiencing it themselves.
Someone is going to build a worldclass “Brain” for enterprises & make a stupid amount of money.
Why? As @da_fant said, “coding w ai is solved bc all context is in the git repo. knowledge work is difficult bc context is spread out. an ai system that creates a git repo w all context for a knowledge worker will be able to 100% automate the work.”
When companies talk about being data ready for AI, this is what they’re implicitly saying.
Engineering has been prepared for this moment for a long time because of the deterministic nature of code, the centralization/versioning of data (read: GitHub), and AI tools that are largely build by engineers for engineers.
But for the rest of white collar work, there’s a TON of catching up to do to properly harness the power of the technology.
The big challenge here, and why no one has truly cracked the code for "an ai system that creates a git repo w all context for a knowledge worker" is because unlike code, most knowledge is 1) distributed, 2) unstructured, and 3) unverifiable.
It's distributed: transcripts live in Granola. Documents in Notion. Customer Data in Hubspot. ERP. Emails. Slack messages. Random spreadsheets. SOP docs. Etc. Etc.
Building an ingestion engine that connects to all of your disparate data sources and auto-updates based on the shelf-life of the data is the first, and frankly, easiest step of the process.
Next, it's unstructured: let's say I want to create a proposal for a potential client. To nail the proposal, I want it to pull important information from a variety of sources. The specific asks & background from our initial sales call. Previous proposals to anchor ourselves to a proven format. And completed sprint boards from Linear, so the pricing & timeline in the document is grounded in truth.
Whether it's a thoughtful filesystem (a la Obsidian) or an OpenClaw-esque memory structure, the brain needs to be great at self-organizing in a thoughtful schema. This is very hard, especially if you want to build a generalizable brain that can be shaped to an array of different enterprises.
And finally, most knowledge is unverifiable: writing a function, running a unit test, and seeing if the code works is easy. It works or it doesn't. Using AI to accelerate your content creation process is highly subjective. What is a good/bad idea? Is the content in your voice or not? Does it feel like slop or novel? Answering these questions are both difficult and non-verifiable.
That same system described above doesn't just have to be great at organizing & forming coherent relationships, but it also has to be great at self-improving based on feedback from the user. Memory systems (like those introduced by OpenClaw) are great to a point, but as you scale the corpus of data within your company's brain, things like compaction and cleaning become wildly important to avoid the needle in the haystack problem.
Someone is going to figure out how to solve this problem, and when they do, not only will they make a shit ton of money, but they'll be robinhood for knowledge workers, enabling non-engineers to enjoy the sort of leverage that only technical folks have felt for the last few years.
62
We built our launch video in Claude Code using HyperFrames.
Now it's yours.
Open source, agent-native framework. HTML to MP4.
$ npx skills add heygen-com/hyperframes
RT + Comment "HyperFrames" to get the full source code of this launch video (must follow)
Doing this for my operations. Will share a video demo soon.
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.
57
This is my vision for AI
claude cowork is making me think maybe we’ll look back and it’ll be obvious that humans were never meant to spend their lives working behind a screen. we’ll see it as inevitable that computers do everything for us on computers and the future of work is cooler than we can imagine
64




















