Format Video / YouTube Embed "Medium CEO Tony Stubblebine criticised OpenAI and Anthropic’s treatment of online writing as undermining consent, fair exchange, and reciprocity between creators and AI companies." Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title Medium CEO Tony Stubblebine: Here's How Writing Evolves In The AI Era Source authors Alex Kantrowitz, Tony Stubblebine Source year 2025 Confidence high Importance high
Format Online paper / PDF "Modern multimodal neural models also process images, audio, and video, whereas text-only language models learn primarily from recorded linguistic material." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title GPT-4o System Card Source authors OpenAI Source year 2024 Confidence high Importance mid
Format Online paper / PDF "Anthropic downloaded millions of pirated books, assembling a digital library from which books were selected for training its language models." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title ORDER ON FAIR USE Source authors U.S. District Court for the Northern District of California, Judge William Alsup Source year 2025 Confidence high Importance high
Format Web Page "OpenAI trains its foundation models partly on information that is publicly available on the internet, alongside other data sources." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title How ChatGPT and our foundation models are developed Source authors OpenAI Source year 2025 Confidence high Importance high
Format Web Page "Text-only language models learn from incomplete records of the world, inheriting some of the missing information and biases contained in those records." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title Improving language understanding with unsupervised learning Source authors Alec Radford Source year 2018 Confidence high Importance high
Format Online paper / PDF "Large language models can absorb linguistic artifacts such as GitHub material, public legal records, Stack Exchange discussions, books, and other recorded human knowledge." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title The Pile: An 800GB Dataset of Diverse Text for Language Modeling Source authors Leo Gao et al. Source year 2020 Confidence high Importance high
Format Online paper / PDF "LLMs took the opposite bet from Cyc, favouring patterns learned from large datasets over the direct encoding of formal commonsense knowledge." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title Getting from Generative AI to Trustworthy AI: What LLMs might learn from Cyc Source authors Doug Lenat, Gary Marcus Source year 2023 Confidence high Importance high
Format Online paper / PDF "With sufficient scale, text synthesis began to look suspiciously like human understanding, while language models demonstrated abilities across different tasks." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title Language Models are Few-Shot Learners Source authors Tom B. Brown et al. Source year 2020 Confidence high Importance high
Format Online paper / PDF "Once scaled with enough data and compute, Transformer language models demonstrated substantial performance improvements." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title Scaling Laws for Neural Language Models Source authors Jared Kaplan et al. Source year 2020 Confidence high Importance high
Format Web Page "The 2017 Transformer breakthrough provided part of the technical foundation for the post-pandemic language-model boom, including ChatGPT’s public introduction in November 2022." View Source Location PART 3 - THE THIRD AXIS CHAPTER 14: AGI & THE SYNTHESIZER Language, in Contrast to Cyc Source title Introducing ChatGPT Source authors OpenAI Source year 2022 Confidence high Importance mid
