Leave your feedback Share Copy URL https://eevb.net/video/2FSYOUPFpk7.html Email Facebook Twitter LinkedIn Pinterest Tumblr Share on Facebook Share on Twitter The Autocomplete Bubble Zheng Qinwen [l92vfPVEoWR] Health Updated on August 06, 2026 EDT — Published on August 06, 2026 EDT Tag: #Zheng Qinwen, #nintendo switch 2 price increase, #stephen kalyn, #cagliari vs udineseThe Autocomplete Bubble Episode 1I've been thinking about getting on here and talking about AI for a while. Everything's coming to a head now, and it's absolutely ridiculous. So we gotta talk about it.Kimi K3 is here. Moonshot's new model is beating Anthropic on benchmarks, it's open source, and if you have the hardware you can download it believe me itv and run it yourself or even sell access to it. Meanwhile, people are acting like the government needs to step in and save American AI companies from better, cheaper software?Let's talk about what happens when a chatbot gets good enough that nobody wants to pay $200 a month for the "frontier" version anymore. Because LLMs are next-token boreham wood vs rochdale predictors. Very, very good autocomplete. Not reasoning engines. Not a path to AGI. And the half-trillion dollar economy built on top of them is looking increasingly shaky.Sources & References:Kimi K3 tops Frontend Code Arena: Reuters Moonshot pauses carrier strike group subscriptions after demand surge: AlphaSignal Beating Claude for 4.6x less money: Axios Trump admin considering bans on Chinese AI models: Tom's Hardware The 2.8T parameter model and the open-weights debate: Yann LeCun departs Meta (Nov 2025): Anthropic x SpaceX Colossus 1 compute deal: Some market figures cited are based on analyst estimates and social reporting. What this channel is: AI hype, debunked. We're not impressed by your parameter count.
Tag: #Zheng Qinwen, #nintendo switch 2 price increase, #stephen kalyn, #cagliari vs udineseThe Autocomplete Bubble Episode 1I've been thinking about getting on here and talking about AI for a while. Everything's coming to a head now, and it's absolutely ridiculous. So we gotta talk about it.Kimi K3 is here. Moonshot's new model is beating Anthropic on benchmarks, it's open source, and if you have the hardware you can download it believe me itv and run it yourself or even sell access to it. Meanwhile, people are acting like the government needs to step in and save American AI companies from better, cheaper software?Let's talk about what happens when a chatbot gets good enough that nobody wants to pay $200 a month for the "frontier" version anymore. Because LLMs are next-token boreham wood vs rochdale predictors. Very, very good autocomplete. Not reasoning engines. Not a path to AGI. And the half-trillion dollar economy built on top of them is looking increasingly shaky.Sources & References:Kimi K3 tops Frontend Code Arena: Reuters Moonshot pauses carrier strike group subscriptions after demand surge: AlphaSignal Beating Claude for 4.6x less money: Axios Trump admin considering bans on Chinese AI models: Tom's Hardware The 2.8T parameter model and the open-weights debate: Yann LeCun departs Meta (Nov 2025): Anthropic x SpaceX Colossus 1 compute deal: Some market figures cited are based on analyst estimates and social reporting. What this channel is: AI hype, debunked. We're not impressed by your parameter count.