GPT (Generative Pre-trained Transformer)
Class of large language models that use transformer neural network architecture; GPT-2, GPT-3, GPT-3.5, and GPT-4 represent successive generations with exponentially more compute and capability
As defined in eJf6QPN9yic
Where it comes up · 10
- Sam Altman and Greg Brockman on AI and the Future (Full Audio)16 claims
- 10 years of NLP history explained in 50 concepts | From Word2Vec, RNNs to GPT15 claims
- The Future of Artificial Intelligence14 claims
- Tyler Cowen on how AI will reorder economies, schools, and spirituality8 claims
- AI Trends 2025: Mary Meeker Deck Deep Dive Part 16 claims
- Bill Gates Reveals Superhuman AI Prediction6 claims
- How ChatGPT changed everything: Thoughts from the frontline of the AI/ML revolution6 claims
- How LLMs Took Over The World4 claims
and 2 more
Most important claims · 10 of 80Search all claims for “GPT (Generative Pre-trained Transformer)” →
- One of Mitchell's biggest fears is that humans will trust AI systems with tasks they are not capable or robust enough to handle, as shown in Harvard Business School research where consultants using GPT-4 performed worse on quantitative reasoning tasks than those not using AI, because they over-relied on the system and made more mistakes.The Future of Artificial Intelligence
- So far, LLMs had been trained using web data which gave them solid understanding of language semantics, syntax, and generation of fluent sentences, but GPT-3's objective was to predict likely sentences with no notion of correctness or factual bias, whereas the human goal in using an LLM is to generate coherent, informative, trustworthy, and reliable outputs.10 years of NLP history explained in 50 concepts | From Word2Vec, RNNs to GPT
- Progress in AI comes from three sources—compute, data, and algorithms—all of which have been progressing at smooth exponential rates across the industry, and technological progress toward AI is not driven by any one company but is a project of all of humanity combining GPUs, algorithmic ideas, and data processing systems.Sam Altman and Greg Brockman on AI and the Future (Full Audio)
- AI systems trained on data from society full of biases will magnify those biases; facial recognition systems are much worse on people of color resulting in false arrests; chatbots trained on internet data produce racist information; and image generation systems cannot create images of Black African doctors treating white kids because it's too far from training data patterns.The Future of Artificial Intelligence
- In high school, Gates was intrigued by AI capabilities like Shakey the robot at Stanford Research Institute, which could engage in reasoning and create execution plans; he initially believed speech recognition and image recognition would be fairly solvable problems, but the Holy Grail was the ability to read and represent knowledge like humans did—something nothing was good at until GPT-4.Bill Gates Reveals Superhuman AI Prediction
- GPT-4 enables semantic understanding and capabilities to emerge from syntactic tasks—as demonstrated in the 2017 'Unsupervised Sentiment Neuron' paper where a model trained to predict the next character in Amazon reviews learned state-of-the-art sentiment classification without explicit instruction.Sam Altman and Greg Brockman on AI and the Future (Full Audio)
- ChatGPT has achieved faster user adoption than any previous technology, reaching 100 million users in 2 months compared to the internet taking 11 years to reach equivalent penetration, indicating AI is on track for horizontal technology status similar to the internet.AI Trends 2025: Mary Meeker Deck Deep Dive Part 1
- GPT models trained at scale with simple objectives (next-token prediction) can accomplish creative tasks like writing sonnets, generating poetry, and producing stylistically varied content that were long assumed to require human-level creativity and semantic understandingHow ChatGPT changed everything: Thoughts from the frontline of the AI/ML revolution
- AI is already smarter than Tyler Cowen in economics; if there were an economics Olympics, GPT-03 would beat him today, not hypothetically, and this is notable precisely because it matters so little to his actual demand for services—his clients want him more, not less.Tyler Cowen on how AI will reorder economies, schools, and spirituality
- Fine-tuning models through tools like LoRA (Low-Rank Adaptation) and access to open-source models allows individuals to customize large language models on consumer-grade hardware, creating a tension between democratic access and safety risks; companies and people worry this could enable malicious actors to fine-tune models for dangerous purposes.The Future of Artificial Intelligence