Benefits of Open-Source AI Models

While happy for OpenAI’s o3 , I’ve decided to end my OpenAI Pro subscription immediately and move 100% to open-source models like (our own) Hominis and DeepSeek. Here’s why: 1. Transparency Over Opacity: Open-source models allow anyone to inspect, modify, and improve their code. This transparency builds trust, fosters accountability, and ensures there’s no "black box" governing how decisions are made—a critical factor in ethical AI. 2. Community-Driven Innovation: Proprietary models are shaped by corporate priorities, but open-source projects thrive on collaboration. By supporting open-source, I’m investing in collective progress over centralized control, empowering developers worldwide to push boundaries equitably. 3. Customization Without Limits: Closed systems often restrict how tools can be adapted. With open-source, I can tailor models to my specific needs, whether for creativity, research, or problem-solving—without waiting for a corporation’s permission or roadmap. 4. Ethical Independence: Relying on a single company’s AI ecosystem risks amplifying its biases, limitations, or profit-driven motives. Open-source alternatives decentralize power, ensuring technology evolves to serve *people*, not shareholders. 5. Long-Term Sustainability: Subscription models lock users into recurring costs, while open-source projects like DeepSeek prioritize accessibility and user agency. I’d rather support frameworks that democratize AI’s benefits, not gatekeep them. This shift isn’t just about tools—it’s a future where technology belongs to everyone.

For years, the open-source community has challenged the closed-source dominance of players. Today, OpenAI has released gpt-oss-120b and gpt-oss-20b, two new open-weight reasoning models. This is a monumental shift, and here’s why it's a game-changer for the entire industry: -Open License: These models come with a permissive Apache 2.0 license, allowing for free commercial use without restrictions—a direct response to developer demand for freedom. -Agentic Power: Built for advanced agentic tasks like tool use and code execution, they're not just powerful but practical for real-world applications. -Deep Customization: They support full-parameter fine-tuning, giving developers unprecedented control to adapt the models to any use case. -Unprecedented Transparency: For the first time, you get full access to the chain-of-thought for easier debugging and higher trust in model outputs. OpenAI's entry into the open-weight space is a major catalyst for the entire AI ecosystem, promising to - Accelerate Competition: This forces all players to innovate faster, release better models, and offer more compelling features to attract developers. The competition drives rapid improvement across the board. - Democratisation of AI: The availability of powerful, open-weight models lowers the barrier to entry for developers and startups. They no longer need multi-billion dollar budgets to access advanced AI capabilities. This enables a wider range of individuals and small teams to experiment, build, and deploy AI solutions, leading to a much larger pool of innovators. -Rapid Customization and Specialization: Open-weight models are perfect for fine-tuning with specific data. Developers can take a strong base model like gpt-oss-20b and specialize it for a niche industry, a company's internal knowledge base, or a unique application. This speeds up the development cycle for tailored AI solutions that were previously too expensive or complex to build. -Community-Driven Development: The principles of open source mean that a global community can now inspect, debug, and improve these models. The LLM market is projected to be worth over $80 billion by 2033, and the fight for developer mindshare is at its core. In essence, this movement can act as a catalyst for the AI landscape to a decentralized ecosystem where innovation can flourish at all levels. 👇 Try them here: - Blog: https://lnkd.in/g4kprY4v - GitHub: https://lnkd.in/gHf2M3mV - Hugging Face: https://lnkd.in/gWESjjDt - Try the models : https://www.gpt-oss.com/ What does this mean for other open models? Let's discuss! 👇

Recently, DeepSeek AI Open-Sourced AI - It Changes Everything. Why? I’ve been building in tech for decades. I’ve seen trends come and go, witnessed the rise (and fall) of hyped-up technologies. But every once in a while, something shifts in a way that fundamentally changes the game. DeepSeek just made that move. They open-sourced their R1 AI model. And if you’ve ever tried to build something with AI, you know why this is massive. For years, the best AI models have been locked away—powerful, yes, but only accessible to those who could afford to pay, play by the rules & operate within the limits set by someone else. Want to tweak the model? Good luck. Want to truly understand how it works? Not happening. That’s why DeepSeek’s decision isn’t just about releasing a model. It’s about unlocking possibility. 𝐖𝐡𝐚𝐭 𝐓𝐡𝐢𝐬 𝐌𝐞𝐚𝐧𝐬 𝐟𝐨𝐫 𝐁𝐮𝐢𝐥𝐝𝐞𝐫𝐬 𝐋𝐢𝐤𝐞 𝐔𝐬 1/ Freedom to Innovate – No more waiting for API updates or praying for access. Developers, researchers, and startups can now build, refine, and push AI forward—on their own terms. 2/ No More Black-Box AI – I’ve lost count of how many times I’ve seen AI models making decisions that no one could explain. With open-source, we can audit, test, and actually trust the tech we build on. 3/ A Level Playing Field – For too long, AI has been a playground for giants. Now, whether you’re a solo founder, a garage startup, or a research lab with a bold idea, you have the same access to world-class AI as the biggest players. 4/ More Efficient, Smarter AI – DeepSeek’s R1 model isn’t just powerful—it’s resource-efficient. This means we can build AI-driven products without needing an army of GPUs or a war chest of funding. ------- Of course, companies in the West may have concerns around compliance, data security, and governance when adopting a foreign AI model. But here’s where things get interesting—DeepSeek isn’t just a model; it’s a technical blueprint. It shows us how world-class AI can be built efficiently. It gives us a roadmap for creating our own models at a fraction of the traditional cost. That’s the real opportunity. Open-source AI isn’t just about making models available—it’s about reshaping the future of how we build. If history has taught me anything, it’s that the best ideas rarely come from closed-door boardrooms. They come from unexpected places, from people tinkering, experimenting, pushing boundaries. Exciting times coming! #deepseek #ai #opensource

No hype -- just facts. 😊 Spent the morning pouring over the GPT-OSS technical report and here's what I've got. OpenAI just released two open source (Apache 2.0) Mixture of Experts (MoE) reasoning models trained for tool use: gpt-oss-20b and gpt-oss-120b. What makes these models special? • They're fully open-weight models with performance similar to paid models like o3-mini and 04-mini • The 20B can run on edge devices and consumer hardware • Both support a massive 130k+ token context length • MoE architecture that makes them efficient despite their size • Strong partnerships with deployment platforms and optimized for compute hardware These models are designed for use in agentic workflows with strong reasoning, tool use, and instruction-following capabilities. You can adjust the reasoning level (low, medium, high) to balance speed vs. depth of analysis. The tool use capabilities are particularly impressive - the models can: • Browse the web to fetch current information • Execute Python code in a Jupyter notebook environment • Call custom functions that you define Performance-wise, gpt-oss-120b actually exceeds OpenAI o3-mini on standard benchmarks like MMLU, GPQA, and coding tasks. Even the smaller 20B model performs surprisingly well despite being 6x smaller than its larger sibling. The models use a special "harmony chat format" that enables advanced features like interleaving tool calls within reasoning steps. The gpt-oss models work out-of-the-box with hardware and deployment providers, thanks to several key partnerships. Fine-tunable with: Hugging Face, Unsloth AI, LLaMA-Factory, Ludwig Deployable with: Hugging Face, Ollama, vLLM, Llama.cpp, OpenRouter, LM Studio, Fireworks AI, Baseten, Vercel, Databricks, Azure, Amazon Web Services (AWS) Optimized for: NVIDIA, AMD, Groq, Cerebras Systems For more details, especially on the training process, adversarial testing, and model performance, check out the blog post or model card. 🔗 Blog: https://lnkd.in/geapnGDE 📄 Model card: https://lnkd.in/gFnYuTUT

The White House just drew a bright red circle around open-source AI. America’s new AI plan says “open-weight models” are the gold standard. That single line changes the game for every builder, student, and small business. Here is why it matters 👇 • Freedom to tinker Startups can download the code, tweak it in a weekend, and launch without begging a vendor. • Data stays home Hospitals, law firms, and city halls keep sensitive files off closed clouds. • Real science Researchers get the weights, rerun the tests, and prove results instead of trusting press releases. • Global signal Open models stamped “Made in the USA” can set the rules before rivals do. • Choice stays local The plan is clear, the dev decides closed or open, the government’s job is to clear the runway. If Washington follows through, the next wave of GPT-level breakthroughs could come from a garage in Ohio as easily as a Big Tech campus. Open weights are more than code, they are economic gravity. Once they roll out, everything pulls toward lower cost, higher trust, faster progress. The question now is simple: Will we sprint into that future or watch someone else claim it first? If you love Open Source AI, support Hugging Face and Mistral AI #OpenSourceAI

OpenAI has just launched two open-source models: gpt-oss-120b and gpt-oss-20b. 𝙀𝙫𝙖𝙡𝙪𝙖𝙩𝙞𝙤𝙣 𝙍𝙚𝙨𝙪𝙡𝙩𝙨 𝘨𝘱𝘵-𝘰𝘴𝘴-120𝘣 𝘰𝘶𝘵𝘱𝘦𝘳𝘧𝘰𝘳𝘮𝘴 𝘖𝘱𝘦𝘯𝘈𝘐 𝘰3‑𝘮𝘪𝘯𝘪 𝘢𝘯𝘥 𝘮𝘢𝘵𝘤𝘩𝘦𝘴 𝘰𝘳 𝘦𝘹𝘤𝘦𝘦𝘥𝘴 𝘖𝘱𝘦𝘯𝘈𝘐 𝘰4-𝘮𝘪𝘯𝘪 on competition coding (Codeforces), general problem solving (MMLU and HLE), and tool calling (TauBench). 𝘾𝙤𝙢𝙥𝙖𝙧𝙞𝙨𝙤𝙣 𝙖𝙜𝙖𝙞𝙣𝙨𝙩 𝙈𝙤𝙙𝙚𝙡-𝙖𝙨-𝙖-𝙎𝙚𝙧𝙫𝙞𝙘𝙚 𝙂𝙋𝙏-4.1/4𝙤 🔶𝘿𝙖𝙩𝙖 𝙋𝙧𝙞𝙫𝙖𝙘𝙮 𝙖𝙣𝙙 𝙍𝙚𝙜𝙪𝙡𝙖𝙩𝙞𝙤𝙣 𝘾𝙤𝙣𝙘𝙚𝙧𝙣𝙨 – Use cases that have strict regulation of data staying within the perimeter can leverage these models either in On-Prem or in the cloud’s dedicated environment instead of the vendor’s hosted platform and leveraging API. 🔶𝙁𝙞𝙣𝙚-𝙩𝙪𝙣𝙞𝙣𝙜 – Full model weights provide a great opportunity to fine-tune full parameters or PEFT for custom requirements. 🔶𝘾𝙤𝙨𝙩– Shifts from pay-as-you-go token cost to compute cost and provides flexibility for R&D. 🔶𝙏𝙧𝙖𝙘𝙚𝙖𝙗𝙞𝙡𝙞𝙩𝙮 – Explicit reasoning visibility level makes it a great option for troubleshooting and tracing model reasoning, and allows an option to fine-tune the prompt & context in agent building.

OpenAI just released open source weights for two of their models: * gpt-oss-120b — for production, general purpose, high reasoning use cases that fits into a single H100 GPU (117B parameters with 5.1B active parameters) * gpt-oss-20b — for lower latency, and local or specialized use cases (21B parameters with 3.6B active parameters) Here are the highlights: * Permissive Apache 2.0 license: Build freely without copyleft restrictions or patent risk—ideal for experimentation, customization, and commercial deployment.  * Configurable reasoning effort: Easily adjust the reasoning effort (low, medium, high) based on your specific use case and latency needs.  * Full chain-of-thought: Gain complete access to the model’s reasoning process, facilitating easier debugging and increased trust in outputs. It’s not intended to be shown to end users.  * Fine-tunable: Fully customize models to your specific use case through parameter fine-tuning. * Agentic capabilities: Use the models’ native capabilities for function calling, web browsing, Python code execution, and Structured Outputs. * Native MXFP4 quantization: The models are trained with native MXFP4 precision for the MoE layer, making gpt-oss-120b run on a single H100 GPU and the gpt-oss-20b model run within 16GB of memory.

The administration released its AI Action Plan today - making a strong endorsement of open source AI and open weights (page 4). I encourage you to read it and understand the critical role open source will play in the future of our country's AI strategy. Don't discount the role that efforts like OpenFL will have in bringing federated learning to government and helping them partner with industry to build the best models available. "Open-source and open-weight AI models are made freely available by developers for anyone in the world to download and modify. Models distributed this way have unique value for innovation because startups can use them flexibly without being dependent on a closed model provider. They also benefit commercial and government adoption of AI because many businesses and governments have sensitive data that they cannot send to closed model vendors. And they are essential for academic research, which often relies on access to the weights and training data of a model to perform scientifically rigorous experiments. We need to ensure America has leading open models founded on American values. Opensource and open-weight models could become global standards in some areas of business and in academic research worldwide. For that reason, they also have geostrategic value. While the decision of whether and how to release an open or closed model is fundamentally up to the developer, the Federal government should create a supportive environment for open models." https://lnkd.in/eHDvT8r3