Jensen称AI模型开发成本不可持续
Jensen has the best gig on the planet. But it is impossible to see how all this is sustainable. Cons...
Gary Marcus分析AI模型开发成本与收益的不平衡，Astra在多项基准测试中表现亮眼。
Gary Marcus指出AI模型开发成本持续攀升，而性能提升却仅是增量式。Epoch AI报告显示Astra模型在ECI基准测试中得分为169，较之前最佳成绩163有显著提升，但仍处于推理时代ECI趋势的不确定性范围内。Astra还在数学、持续学习和游戏谜题基准测试中创造了新记录。
Jensen has the best gig on the planet. But it is impossible to see how all this is sustainable. Cons...
Jensen has the best gig on the planet. But it is impossible to see how all this is sustainable. Consider the following: Each new model costs vastly more to develop. That’s fabulous for him! On the other hand, each new model offers only incremental gains. And OpenAI and Anthropic need to keep cutting costs, dramatically, with every release. This is truly great for consumers! But I cannot see how on earth the model providers are ever going to become sustainably profitable, given all those facts. Epoch AI @EpochAIResearch GPT-6 Astra has set a new ECI record, with a score of 169. This is a substantial jump from the prior best (163), but is within our uncertainty range for the reasoning-era ECI trend. Astra also set new records on our math, continual learning, and game-puzzles benchmarks. On our long-horizon coding benchmark, MirrorCode, Astra ranks between Opus 4.7 and Fable 5. OpenAI gave us pre-release access to test Astra. Charts and more details for Astra’s individual benchmark results in the thread. 🔗 View Quoted Tweet 💬 0 🔄 0 ❤️ 2 👀 583 ⚡
