Raakin ROll 🇦🇪🇺🇸 - راكين رول

🚨 Are we seeing early signs of AEI? We did a live comparison between ChatGPT and a DXB Chatbot linked to our Engine. You will see how the responses took into account the context and started suggestions based on the general knowledge of the initial inquiry. This allows the system to really start to expand. For instance, in the video you will see when asked about "mobility" options, ChatGPT (with Web access) (or any standard or intelligence LLM) started to make suggestions..WHEREAS Nucleus AI engine made suggestions that were near each other given "Mobility" was a issue. Nucleus AI will continue to work towards AGI and advance #AI but was excited to start seeing this develop. ....For the record, we are NOT building an LLM.

@michael_chomsky Appreciate the vision here but this isn't an idea.md anymore — We've been shipping this for 2 years. filed two PCT patents on the memory architecture. happy to show you how it works

Here’s an idea.md for anyone who isn’t scared to build an ambitious product: Someone’s going to make 100-1000M dollars building a self-updating personal knowledge base that syncs with imessage, twitter, email, chatgpt/claude/claude code/codex messages. This knowledge base will have an MCP to be accessible from anywhere. You’ll be able to edit it like notion, style it to your liking, and set rules about how data is organized. Once it gets mature enough, it can even proactively suggest things that will improve your life, as it knows everything about you. Unlike some memory systems, this will just be files so completely observable. The closest thing to this that exists is Mintlify’s KB and Notion, but both are more for enterprises than individuals. Just paste this into Claude Code, ask it to perform a socratic interview about ambiguities (or just use gstack), clank for 2 days, get Chamath and Karpathy as your first two customers, and do YC. You’ll have competition at some point but it won’t matter because you’ll be better at UGC and paid ads.

@chamath We are doing exactly that with Nucleus. @nucleusagi . Not even just just sync your chats but using our two filed patents hierarchical memory systems.

This may be a dumb question but I’ll ask it here anyways: I can’t find a good way for my various AI chats to automatically sync its conversation history into a structured knowledge base. So that as I update various chats from time to time and refine context, my knowledge base automatically grows with this new info.

A lot of noise today about Pakistan owing the UAE gratitude. Worth some perspective. Pakistan has deployed military trainers and defence personnel who have been part of the UAE's defence architecture for decades, trained Emirati pilots, and contributed institutionally to the country's armed forces. PIA leased aircraft and supplied pilots to Emirates airline when it launched, helping build what is now one of the world's largest carriers. Pakistani professionals helped establish the UAE's early banking infrastructure. Pakistanis are among the largest business owner and investor communities in the country, contributing enormously to its commercial fabric across construction, trade, retail, and services. The UAE in return has provided Pakistan with financial deposits during difficult periods. That is real and it is appreciated. But deposits placed at interest, now recalled, are financial instruments. They are not charity. Millions of Pakistanis live and work in the UAE under conditions that often offer fewer rights and protections than those extended to Western expatriates. Visa access to the UAE for Pakistani nationals has been severely restricted for years. These are facts that tend not to feature in the narrative of UAE generosity. This is a long relationship built on mutual benefit. Pakistan gave the UAE its military trainers, its aviation expertise, its banking talent, its labour, and its diaspora capital. The UAE gave Pakistan financial support and employment. Neither side has been doing the other a favour. They have been doing business. Framing it as Pakistani ingratitude reflects a poor understanding of what this relationship has actually been.

Thousands of people on X paste headlines into Grok every day asking "is this true" and get a confident answer from a model that has no idea. That's not verification. that's pattern matching with good grammar. This is what actual verification infrastructure looks like at scale.

She absolutely deserves it! #Swifty

Hot take: AI models are like fresh food. When a new model drops, it's incredible. Best answers you've ever seen. Then after a few weeks it gets sluggish. Hallucinating more. Conversations don't make sense. So you switch to the next model. And start over from scratch. That's the problem. Models change. Your context shouldn't have to.

Hot take: AI models are like fresh food. When a new model drops, it's incredible. Best answers you've ever seen. Then after a few weeks it gets sluggish. Hallucinating more. Conversations don't make sense. So you switch to the next model. And start over from scratch. That's the problem. Models change. Your context shouldn't have to.

Your pitch deck changes every week. Your AI doesn’t know that. We built AI Instructions in Nucleus — one line that says “this is a living document, check for newer versions before referencing.” Replaces fine-tuning. No retraining. No compute cost. Context that evolves with your business.

Someone said this wisey- "I see your Taco and I raise you NACHO, Netanyahu Always Causes Hellish Outcomes. Meaning to say that even if trump chickens out, Netanyahu can still force his hand or actually bomb infrastructure to trigger the chain reaction regardless and make sure the situation does not deescalate."

@zanzaf That's exactly it! Most people think the model is the product. The context is the product. The model is just the renderer.

@Raakin Nucleus (Persistant) Context

Working on something that I haven't seen anyone else do yet. Saved a full conversation from Claude into my Nucleus AI brain. Opened ChatGPT. Asked it to build on that work. It pulled everything. The pitch strategy. The context. All of it. Automatically through our MCP connector. Models change. Your context shouldn't have to. Showing this properly soon.

@DanQayyum YOLO

The re-explaining problem is real but persistent memory alone doesn't fix it if the memory itself is unreliable — conflicting sources, stale data, no provenance — you've just made the agent more confidently wrong across more sessions identity layer + verified context layer is the full picture. one without the other is still broken

persistent memory for agents is the unlock sovereign context layer on MCP infrastructure solves the re-explaining problem that kills every serious AI workflow worldcoin agentkit, base skills, virtuals on bnb... everyone building identity and verification layers right now because agents need to remember who they are across sessions mnemos hitting the right problem at the right time

The demo explained🧠 We recorded a simple demo. We tell Claude we prefer iOS over Android. Then go back to the MNEMOS vault and watch that preference appear as a new memory — automatically, in real time, without touching anything. That sounds small. It isn't. What you're watching is the end of a problem that every serious AI user knows: the constant re-explaining. Every new session, every new tool, every new model — you start from zero. Your preferences, your context, your history, gone. You become a stranger to your own tools. What MCP + MNEMOS does is build a persistent layer beneath all of it. Your AI doesn't need to remember you because your vault does. And your vault is yours — encrypted, sovereign, not dependent on any single product or company staying in business or keeping your data intact. No manual input. No copy paste. No starting over. Your memory, finally working for you. 🔗 mnemosbase.com

Something I keep thinking about. When you use Claude through Perplexity, you're not really using Claude. They've modified the base prompts. Tuned it for their platform. Same with every aggregator. It's the Samsung problem. You buy Android but you're getting their version of it. Modified. Bloatware added. I prefer going direct. Claude through Claude. GPT through GPT. Raw models. That's why we built Nucleus as a connector layer, not a model wrapper. Your Claude stays your Claude. We just give it Contextual Memory. "Models Change. But your Context Remains"

@EmmanuelMacron At least a Verb?

🇫🇷

@alialsalim @ianbremmer We still have the next 10 days after for the extra credit. Runway secured

@ianbremmer Its ok, most of the region is currently fasting

90% of drinking water in oman, qatar, and kuwait comes from desalination plants. bahrain already had one struck by a drone. jebel ali port — 12 miles from one of the world's largest desalination facilities — has also been hit.

The AFP article on AI dashboards said the creator "vibe-coded" World Monitor in a weekend using Claude Code. That's incredible for shipping speed. But the analyst in the same article said "the hallucination risk is real" when AI determines what information is significant. Speed and accuracy are two different problems. Most people are only solving for one.

@AkinwoleKunel @officialladi_T You should check out the context demo by @nucleusagi and basically had Opus 4.6 w/ Web Search vs. Opus 4.6 + Nucleus .. made such a day and night difference

@officialladi_T Ask Chatgpt and claude both will explain depth in for you

Don't know if you will make it as a beginner in data analysis?? Just start that course and don't overthink it Clarity comes after an action is taken and never before Anyhow just start 🔥

the move makes sense — and it gets even better when you fix the context layer underneath it not talking about window size, that's a different thing. talking about what claude actually reasons over — whether the information it's given is reliable, non-contradictory, and epistemically sound before it even starts thinking that's where the real ceiling is and most teams haven't hit it yet

If I could short OpenAI, I would. We just cancelled GPT for the entire Hampton team. Claude replaced it, and added crazy functionality that GPT cant match. Coincidence that OpenAI is now running ads and chasing enterprise contracts? They know.