'We're making exponentially more internet every day': researcher pushes back on pretraining data exhaustion claims
ZeeshanZiaML · x · 2026-10-19
Pushing back on the popular claim that the internet has been 'used up' for pretraining, the author argues we're generating exponentially more internet content every day — much of it now AI-produced. A reply raises the counterpoint: most of this new data may be locked inside walled gardens like Twitter, raising open questions about both crawlability and synthetic data quality.
Related event: Developers Push Back on 'Internet Data Exhaustion' Claim(2 posts)→
More from Models
- SemiAnalysis: Anthropic A/B testing ~20% lower usage limits on some accounts — imjustnewatai · 2026-10-23
- Swift-1.5-Qwen3.8-27B ported to hipfire MQ4: 2.2x faster end-to-end than GGUF on a 7900 XT — Queasy_Asparagus69 · 2026-10-23
- Boundless offers abliterated DeepSeek V4.1 Flash for defensive security testing — MoonL88537 · 2026-10-23
- Anthropic uses Claude to formally verify security properties in 100+ open-source libraries at RISC-V assembly level — AnthropicAI · 2026-10-23
- Are models in an uncanny valley? Researcher warns users to stay vigilant — danielrock · 2026-10-23
- Uncensored Swift 1.5 Qwen 3.8 Flash Next Released as 4-bit GGUF with Strata-Ready Artifacts — jjusko20 · 2026-10-23