Updated 08/18/2025
Intro
Hey, this is my first blog on AI. If you find it interesting, feel free to go check out some of the past weekly newsletters! The goal of this blog will be to serve as a basic guide to understand what companies in the AI space are doing, identify their potential long term goals, and establishing the best AI resources for different tasks.
This will be a dynamic blog. I will update it over time as I try out new AI tools. I’m writing this for folks who are just getting started with AI, or for those who primarily use one or two AI products per week (ChatGPT, CoPilot etc.). It’s not meant to be advanced but rather a nice, easy guide to AI. Ok, let’s go.
Content:
AI Buzzwords
Useful companies by sector
Tabled breakdowns by company (Basics, Current Details, Future)
AI Dictionary
Here are some words I’ve been hearing a lot in the AI space.
| Name | Description | ELI5 |
|---|---|---|
| AGI | Artificial General Intelligence | Like a robot that can do almost any job |
| ASI | Artificial Superintelligence | Smarter than every human on Earth |
| Vibe Coding | Prompting AI to write code for you | You tell it what you want, it builds it |
| LLM | Large Language Model | A giant brain trained to talk |
| Agent | Autonomous AI system | An AI that can go do stuff for you |
| Retrieval (RAG) | Real-time info lookup | It Googles before answering |
| Token | Chunk of a word | AI reads in puzzle pieces, not full words |
| Context Window | How much info an AI can remember | Bigger brain = remembers more of chat |
| Multimodal | Accepts text, images, audio, video | Like a human — it sees and hears things |
| Inference Time | Time it takes for AI to respond | How long it takes to talk back |
| GPU | Graphics Processing Unit | The engine that powers AI thinking |
| Prompt Injection | Sneaky input that breaks the AI’s logic | Like tricking AI with a secret message |
| Foundation Model | A model trained for general tasks | One big model that can do a little of everything |
| Open Source | Code anyone can use or modify | Free and shareable AI for anyone to build with |
| CLI | Command Line Interface | A text box where devs type commands |
| GPT Wrapper | Custom software built around ChatGPT | Makes GPT do specific jobs for you |
| Parallel Thinking | Running multiple ideas at once | AI thinks in a bunch of directions together |
| Embodied AI | AI with a physical form like a robot | Not just smart, but walking around |
| Generative AI | AI that creates things like text, images, or code | AI that makes stuff from scratch |
| Agentic AI | AI that can plan and act on its own | You give it a goal, it figures out the steps |
Some Useful Tools
This is a starting set of AI tools I use and the applications they fit best. The list is small for now but will expand a lot over time.
| Task | Top Tool | Why It Wins |
|---|---|---|
| Writing | ChatGPT (OpenAI) | Fast, flexible, great for drafting and editing all kinds of text |
| Research | Gemini (Google) | Strong search integration and summarization capabilities |
| Video | Gemini (Google) | Veo3 produces the best consumer-facing AI video so far |
| Photo | Midjourney | Jury's not fully out on this but I've seen some good stuff |
| Voice | ElevenLabs | Natural-sounding voice cloning and TTS with emotion control |
| Music | ElevenLabs | Strong option for AI-generated vocals and music-style synthesis |
| AI Avatars | Hedra | Best balance of realism and accessibility for AI-driven avatars |
| Video Editing | Descript | AI editing, auto-captioning, and audio cleanup in one tool |
| Meetings | TL;DV | Clean summaries and transcription for team meetings |
| Coding | Claude (Anthropic) | Great reasoning for code explanation, debugging, and generation |
| Design | Figma / Canva | Figma is excellent; haven't tried its AI integrations but trust they're good. Canva I don't like, but it has some decent AI tools |
| Brainstorming | ChatGPT (OpenAI) | Handles idea generation and iteration really well |
| Task Automation | Zapier (Basic) / Claude Code (Custom) | Zapier connects AI to workflows without coding. Claude Code allows deeper customization |
| Scheduling | Blockit | Streamlined AI-assisted scheduling and calendar coordination |
| Building | Claude Code / Lovable | Idea → App with AI-assisted coding and deployment |
AI Company Baseline
Here’s a basic outline of who I think the top 10 AI companies are right now (mainly consumer-facing, non-specialized). The table lists their main product, parent company, and CEO. I’ll add Monthly Active Users (MAU) once I’m more confident in the numbers.
| Product | Company | CEO |
|---|---|---|
| ChatGPT | OpenAI | Sam Altman |
| Perplexity | Perplexity.ai | Aravind Srinivas |
| Gemini | Sundar Pichai | |
| Claude | Anthropic | Dario Amodei |
| Copilot | GitHub (Microsoft) | Thomas Dohmke |
| Grok | xAI | Elon Musk |
| Cursor | Cursor | Ammar Reshi |
| LLaMA | Meta | Mark Zuckerberg |
| Deepseek | Deepseek | Liang Wenfeng |
| NVIDIA | NVIDIA | Jensen Huang |
AI Company Use Cases
This table includes the same companies alongside their generic use cases, as well as what I would consider their key feature. I’ve italicized Deepseek, Copilot, and Perplexity because I’ve used them a lot less than the others (technically I use NVIDIA everyday by proxy).
| Company | Description | Top Feature |
|---|---|---|
| ChatGPT (OpenAI) | Good for general conversations, research, voice chat, and integrated GPT library/app store. Also has image and video generation (pales in comparison to Gemini). Recent Agent Mode release could be very cool. Excited for upcoming GPT Browser and GPT-5. Great consumer-facing product. | General Chat / Voice Mode |
| Gemini (Google) | General chatbot with strong research capabilities, email summarization, search integration, and seamless integration across Google Workspace tools. Easily the best video generation model (Veo3). Also has solid voice mode. | Veo3 (Video Gen) |
| Claude (Anthropic) | Great for software engineering. Claude feels like a helpful assistant; Claude Code acts like a true programmer on technical tasks. | Writing Code // Claude Code |
| Grok (xAI) | Chatbot with image generation, research, and a growing set of AI personas (like Grok, Ani, and Kids Mode). Best known for tight integration with X and real-time reply summarizations. | Social Media Integration |
| Meta (LLaMA) | Open-source (for now) chatbot with image generation. Still needs some work, but shows potential. Strong base for developers and experiments. | Open Source |
| NVIDIA | Not a chatbot, but the infrastructure giant behind the AI boom. Dominates AI chip production and powers most modern AI models through its GPUs and supercomputers. | AI Hardware / GPU Ecosystem |
| Perplexity | Real-time research and fact-based summarization engine. Great for citations, source-backed answers, and staying up to date with current events. | Search + Sources |
| GitHub Copilot | Embedded across Microsoft apps like Excel, Word, and Outlook. Great for in-line AI help, writing, and data analysis. Especially helpful for working professionals. GitHub Spark, their new text-to-full-stack release, should be intriguing. | Document/Spreadsheet Support |
| Cursor | An AI-powered coding editor that helps write and debug code inside its own environment. Works as a coding assistant rather than a full agent, making it good for quick fixes and guidance while staying integrated. | In-App Coding Assist |
| Deepseek | Chatbot with some programming capability. Strong at debugging. Will update more once I use it more. | Cost-Efficiency Build |
AI Company Future
This table goes a bit deeper, covering my take on each company’s future outlook and their chances of success. It was a fun one to put together, and I’ll update the TBD entries as I get more hands-on experience with those tools.
| Company | Detailed Outlook | Success Likelihood |
|---|---|---|
| NVIDIA | Barring disaster, NVIDIA will have success in AI. There's that old quote: the most successful person in a gold rush is the one selling shovels, and that’s very much the case here. NVIDIA is the clear frontrunner for GPUs, and I don’t see that changing anytime soon. The biggest question for NVIDIA is the same one that applies to AI as a whole: how will energy production, or lack thereof, affect growth? China is currently beating the U.S. in energy, and NVIDIA is facing some pressure from the U.S. government about selling chips to China. If the U.S. runs into energy issues while China accelerates, and demand continues rising, we could see some geopolitical tension. | 98% |
| Gemini (Google) | I’m really intrigued by Gemini’s long-term potential in video, especially how it might integrate with Hollywood. Right now, they clearly have the best video modeling software, at least on the consumer-facing side. There’s also huge potential in advertising, since it’s Google. Long term, I’m really curious to see what “Google Search” looks like in 10 years. Plus, there’s all the potential for integration across Google tools like Docs, Gmail, Android, and more. | 92% |
| OpenAI (ChatGPT) | OpenAI definitely has the edge right now. Even folks who don’t use AI much still use ChatGPT. It’s the number one consumer-facing AI, and so far, they’ve had tons of success. They’re leading the charge, and it’s hard to see them falling off the mountain. That said, there are some concerns. The SoftBank funding talks seem to have fallen through, which raises questions. And Sam Altman remains something of a wildcard. The for-profit vs. nonprofit drama got weird. But I’d like to think I still believe in him. With their consumer lead, OpenAI should be fine as long as they keep pushing forward. I’m especially intrigued to see what happens with GPT Agent, GPT Browser, and continued product development. | 92% |
| Anthropic (Claude) | Claude’s future feels firmly rooted in agentic software engineering. Claude Code already handles small- to mid-size tasks like refactoring, helper logic, and quick builds, but I can see it evolving into a full-stack engineer on its own. The real question is what the cost structure will look like. Will we be paying $10K per year for a Claude-powered dev? Will companies offer “Claude-as-a-coworker” licenses? It’ll obviously vary by use case, but that’s the trajectory I see. They may have some financial issues at the moment, which causes a bit of concern. There was chatter about Apple acquiring Anthropic, their parent company, which I don’t think will happen — but it should. | 90% |
| GitHub Copilot | Copilot seems to be used a lot by people I talk to. It feels like it has a smaller hill to climb given its integrations with GitHub and Microsoft. Used it a bit, solid tool. Once again, given its integrations and its shorter destination it should be just fine. | 88% |
| xAI (Grok) | Even though Grok is currently on par, or even behind, ChatGPT and Gemini in some areas (see: the International Math Olympiad benchmark, where Chat excelled), I think its long-term future lies in math and physics. Elon may stay engaged with the social and consumer-facing side for now, but over time, I see xAI shifting its focus. I believe Grok will eventually become a powerhouse in math, physics, and robotics, with deep integration into Tesla and other Musk ventures. I see it becoming far more useful as an internal engine powering ambitious hardware and science projects. Given Elon’s track record, I think it will succeed. Starting to have some doubts on Grok given its social direction. | 85% |
| Meta (LLaMA) | Meta is spending a massive amount of money and you'd expect that to yield results, assuming AI isn't a bubble. They made a similar bet on AR/VR with the rebrand to Meta, which didn’t pay off (yet), but I still believe in AR/VR long term. They were just early. I’m not completely sure what their AI future looks like, but I believe in the space and they’re throwing hundreds of billions at it. I imagine they’ll find success. Their biggest long-term advantage might come from AR/VR mixed with AI: full virtual realities with real people and AI characters. The advertising opportunity is enormous as well. Leadership seems to be missing from Meta. | 83% |
| Cursor | Cursor is adapting well to the AI landscape. Initially seen as more of an assistant, it has recently added integrations with GPT-5, Claude, and other tools to embed CLI Agents within the platform. Much Improved. | 80% |
| Perplexity | Perplexity seems to be well regarded for information and retrieval. Now that I've used it more, I can confidently say I do not get the hype around it. Deep research, which from my understanding was one of its most touted features, pales in comparison to other tools. Perplexity is nice but far from the competition IMO. | 70% |
| Deepseek | I haven’t used Deepseek enough to provide an accurate prediction on its future yet. The thing that gives it the most value is the fact that it’s achieved a ton of success despite having way less funding and build time than something like ChatGPT. That alone makes it worth watching. | TBD |
That’s all I got. If you saw anything you have input on, feel free to shoot me an email: [email protected]! I’ll plan to keep this updated whenever I find new tools or new information.