Mohammad Arshad (@arshad83) header
Mohammad Arshad (@arshad83) profile picture

Principal Data Scientist | 120K+ Linkedin Community ! Generative AI | 20 Years+ Exp | Ex-Accenture, HP, Dell | Keynote Speaker & Mentor | LLM, AWS, Azure GCP

AI Agents & RAG Tutorials β†’
Joined April 2009
Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)
Mohammad Arshad retweeted
Want to upgrade your Artificial Intelligence game for 2026? We're giving away $2,000+ worth of premium AI resources β€” 100% FREE! What's inside? πŸ“·AI Courses & Program πŸ“·Exclusive Resources & Hackathon πŸ“· Generative AI Courses 1️⃣ Follow Me & @decodingdatasci 2️⃣ Like & Repost this 3️⃣Reply with "Send" Don't miss this if you're serious about mastering AI in 2026!
2
2
2
95
Mohammad Arshad retweeted
We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of the Lord of the Rings, a 1M token budget (~$10) and asked for three js render of it. Opus went off for ~2 hours and wrote 5500 lines of code that (procedurally) rendered the story. It's kind of janky but fun. But it's a bit mindboggling that the LLM has to place and orchestrate various polygon assets in (x,y,z) coordinates and write code that animates it all, and that it even does anything at all. I also like this kind of examples because no one in their right mind would ever spend the time to write something this custom but LLMs have all the stamina and patience in the world, so it's an example where we go from "no one would ever do this" to "sure, why not, it's ~free". There might be a lot more. But I'm excited about creating hyper custom worlds that you can imagine dropping players into, e.g. here to participate in the LoTR story as a spectator NPC, or one of the characters, or etc. Something like an ephemeral GTA of X on demand. Last thought is that the domain of worlds/games exposes a weakness in LLMs: they can't easily audit their work because they aren't able to efficiently and natively perceive videos or play games within them. Here, Opus 5 had to very slowly and painstakingly take screenshots at different points, and it messed up a few times and created a bunch of jank. An example of raw capability (multimodal, gameplay) that I think is still quite lacking.
1,804
2,312
753
29,790
6,321,523
Mohammad Arshad retweeted
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.
4,905
14,050
9,951
103,156
57,750,916
Mohammad Arshad retweeted
Most AI agents fail for one reason: People build them without a system. A good AI agent is not just a prompt. It’s a structured pipeline from idea β†’ execution. Here’s the framework that actually works: Define the job What problem is the agent solving? Who is the end user? What does success actually look like? Design the brain Create a strong system prompt: β€’ Clear role β€’ Instructions β€’ Constraints β€’ Guardrails This layer decides whether the agent feels reliable or broken. Choose the right model Balance: β€’ Intelligence β€’ Speed β€’ Cost Not every task needs the most expensive model. Add tools Connect APIs, databases, MCP servers, and custom functions. Agents become useful when they can take actions, not just generate text. Add memory Use short-term and long-term memory so the agent can retain context, adapt, and improve over time. Orchestrate the workflow Handle: β€’ Triggers β€’ Retries β€’ Multi-agent communication β€’ Workflow logic This is where production systems are built. Build the interface Turn the agent into something people can actually use: β€’ Chat app β€’ API β€’ Slack bot β€’ Internal tool Functionality alone is not enough. Test and iterate Run evals. Measure latency. Collect real user feedback. Iteration is the real competitive advantage. ❀️ Like πŸ” Repost πŸ”– Bookmark Follow @Origin_AI_01 for more AI insights. #AI #ArtificialIntelligence #GenerativeAI #AIAgents #Automation
10
7
38
985
What makes Genspark AI Slides 5.0 interesting is that the tool isn’t just generating presentations it’s embedding expert structure directly into the workflow Slide Skills + Built-in Data Analysis + Custom Team Skills makes Genspark feel much bigger than standard AI slide automation
Introducing Genspark AI Slides 5.0. Our biggest upgrade to AI Slides yet. The biggest leap? Slide Skills. Reusable assets that bundle expert thinking and polished design, ready in one click. Here's everything new in 5.0: - 100+ Expert Skills built in. We've packed the thinking of the world's top experts, across consulting, sales, teaching, and more. Apply with one click, and their expertise becomes yours. - Build your own Skill. Upload a deck or PDF, describe the analysis steps, and lock in your team's style. - 5Γ— faster, higher quality. Powered by next-gen architecture and the most powerful models. - Built-in data analysis. Crunches your Excel and Sheets data into presentation-ready insights. Every deck thinks like an expert. Try it now β†’ genspark.ai/ai_slides πŸŽ‰ 50% off AI Slides 5.0 credits, 14 days only
15
16
69
19,000
Mohammad Arshad retweeted
You can make Claude, ChatGPT, and Gemini 20x smarter… just by adding one sentence These are 10 β€œdecisive” and powerful sentences that no one uses: πŸ‘‡ [ Add to bookmarks πŸ”– you'll need it]
55
57
125
11,705
Don’t PAY APPLE for extra iCloud storage. I fixed it in 10 minutes I freed up 12GB without paying a dollar. Use these 8 simple steps:
7
18
36
10,962
Mohammad Arshad retweeted
Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.
4,578
9,836
6,728
79,436
14,483,439
Mohammad Arshad retweeted
πŸ’₯BREAKING: Seedance 2.0 is now officially available on GlobalGPT at 50% OFF! Realistic physics, native audio-video generation, and best-in-class image control for AI video. Now avaiable for all regions. No limits. No restrictions. No invite codes.πŸ‘‡
🀝 Paid partnership
29
16
52
2,838
Mohammad Arshad retweeted
Adobe and Figma after the great big announcement😩 I hope you all know, who made that??🀫
18
13
36
965
Mohammad Arshad retweeted
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
The degree to which you are awed by AI is perfectly correlated with how much you use AI to code.
1,264
2,704
718
22,758
7,671,480
Mohammad Arshad retweeted
this is wild… 23 minutes. fully AI. And for once, you’re not thinking about the tech, you’re actually feeling the story. Higgs just changed what AI video is supposed to do.
🀝 Paid partnership Β· πŸ€– Made with AI
We just made a 23-MINUTE sci-fi pilot in 4 days. And it is 100% AI.
8
9
39
5,571
Mohammad Arshad retweeted
Wow, this tweet went very viral! I wanted share a possibly slightly improved version of the tweet in an "idea file". The idea of the idea file is that in this era of LLM agents, there is less of a point/need of sharing the specific code/app, you just share the idea, then the other person's agent customizes & builds it for your specific needs. So here's the idea in a gist format: gist.github.com/karpathy/442… You can give this to your agent and it can build you your own LLM wiki and guide you on how to use it etc. It's intentionally kept a little bit abstract/vague because there are so many directions to take this in. And ofc, people can adjust the idea or contribute their own in the Discussion which is cool.
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.
1,121
2,824
685
26,768
7,355,427
Mohammad Arshad retweeted
2026 is coming to an end πŸ₯Ή here are my most used AI Tools This Year: Replit - Build mobile apps Antigravity - Best AI for coding Arcads AI - Marketing for apps Higgsfield AI - Stunning AI videos Gemini 3 - AI image editing Perplexity AI - Web search Lovable - AI Websites Grok-4.1 - Deep research tasks Typefully - Social media manager Gamma AI - Generate presentations So, what was your favourite AI tool in 2026?
30
32
99
3,521
Mohammad Arshad retweeted
$2,000,000. From reposting simple kids videos on YouTube. 🀯 No filming. No face. No complex editing. I documented the exact step-by-step system. Free for the next 24 hours. To get it: 1. Follow @Alaina_Jahan_Ai 2. Like + RT 3.Reply β€œSEND” Clock’s ticking. ⏳
23
13
36
541
Mohammad Arshad retweeted
I tested 4 AI tools with one specific prompt. The Results: β€’ Grok & Claude: Provided high-quality research and text. However, the output remained a document. No slide design was provided. β€’ Gemini: Organized the content into a structured outline, but the final layout lacked visual hierarchy. β€’ AiPPT.com: Converted the prompt directly into a formatted deck. While others generate the content, AiPPT.com builds the presentation. It handles the design, layouts, and visual flow automatically. If you need a finished file rather than a text draft, the choice is clear.
🀝 Paid partnership
26
12
68
39,663
Mohammad Arshad retweeted
🚨 BREAKING: Google Gemini can now analyze stocks like a Wall Street pro for free. Here are 10 game-changing Gemini prompts that replace a $4K/month Bloomberg terminal:
12
5
24
587
Mohammad Arshad retweeted
There’s a moment when using Uni-1 by @lumalabsai where you look at the output and think β€œwait… I didn’t even explain it that well.” But it still got it right. That’s what surprised me the most. It’s honestly INSANE. From rough ideas to visuals that actually make sense. Tried a few different styles and concepts ↓ Prompts ↓ Some of these genuinely feel unreal. Try the model here: lumalabs.ai/app #LumaPartner #ai #contentcreation #tech
πŸ€– Made with AI
22
11
49
487
Mohammad Arshad retweeted
🚨 119𝐆𝐁+ π†πŽπŽπ†π‹π„ πƒπ‘πˆπ•π„ β€” 𝐀𝐋𝐋 ππ€πˆπƒ π‚πŽπ”π‘π’π„π’ 🚨 𝐌𝐒𝐬𝐬𝐞𝐝 𝐒𝐭 π₯𝐚𝐬𝐭 𝐭𝐒𝐦𝐞? πˆβ€™π¦ 𝐝𝐫𝐨𝐩𝐩𝐒𝐧𝐠 𝐒𝐭 𝐚𝐠𝐚𝐒𝐧. πŸ‘Œ This vault helped agencies close $9K+ clients using proven systems. Inside the Drive: πŸ“ AI & Automation πŸ“ Ethical Hacking πŸ“ Cybersecurity πŸ“ Prompt Engineering πŸ“ Google Cloud πŸ“ Machine Learning πŸ“ DevOps + CI/CD πŸ“ Docker & Kubernetes πŸ“ Blockchain πŸ“ Power BI πŸ“ React + Node πŸ“ Cloud Security πŸ“ Linux πŸ“ Pen Testing πŸ“ Data Analytics πŸ“ Data Science πŸ“ Big Data πŸ“ SQL πŸ“ Tableau πŸ“ Python πŸ“ AWS πŸ“ Java Everything organized. Everything premium. 119GB+ value. Get it: βœ” Follow @rosemoni18 [MusT] βœ” Like & RT βœ” Comment β€œ NEED ” To Get Auto DM. ⚠️ No follow = No Access ⚠️
32
26
41
860
Mohammad Arshad retweeted
πŸš€ Turn any video into anythingβ€”instantlyβ€”with APOB ReVideo. Swap characters, styles, scenes, outfits, camera motion, and products in seconds. AI video editing just got limitless. ➑️ Try it free now : mega.apob.ai/MagiberRahman
25
17
1
37
10,956