期权我就看看
2024-10-17
$英伟达(NVDA)$
就一个问题，靠算力堆出来的人工智能，是真的人工智能吗，难道不是更应该靠神经网络算法的不断升级吗？
免责声明：上述内容仅代表发帖人个人观点，不构成本平台的任何投资建议。
分享至
微信
复制链接
精彩评论
Patrick Tsui
2024-10-19
Patrick Tsui
去Google算法和神经网路又不难，别只在网路上见过这两个词就觉得自己懂ai🤡
期权我就看看
:
嗯，还是你牛逼，会用google搜算法理论就懂AI了
老当益壮铁树开花
2024-10-19
老当益壮铁树开花
就跟曾经的超级计算机一样，靠堆cpu实现的算力不及10年后的一颗cpu算力，其实都是资本的阴谋，否则哪来翻倍的想象力。当然很多所谓懂AI的人要嗤之以鼻了。
期权我就看看
:
哈哈，看看你楼下那位不就是吗，就是因为算法还不行效率太低了所以才需要上大算力才能出结果，但是算力是马上花钱买卡能解决的，算法可不是一时半会儿就能有质的飞跃，多数人只能看到眼前的利益
Xzhouz
2024-10-18
Xzhouz
这也是我一直思考的问题，如果堆量可以实现，那么买三五个低端卡代替一个高端卡，能解决这个问题。剩下的无非是成本和能耗的问题。所以制约人工智能的最重要的因素到底是高端显卡，还是算力，还是软件？
二狗大石
:
不冲突啊 算力是基础 神经网络算法也是基础 两者都要更好才能更好 至于3个低端卡换一个高端卡 是完全不划算的 多个低端卡提供高端卡一样的算力 能耗是更高的
二狗大石
:
我跑过一些简单的人工智能模型 例如图片的智能识别 第一 需要合适的算法以及正确的标注 第二 需要高端的显卡才能快速跑完一次学习 如果算力不够 跑一次训练就要很多天 那就效率太低了 如果我一天就能跑完一次 你要100天才能跑完一次 那效率肯定是天壤之别的（跑完以后得到的模型又要不断的测试优化 真实的开发是需要跑无数次）
查看全部7条评论
APP内打开
发表看法
7
4
{"i18n":{"language":"zh_CN"},"detailType":1,"isChannel":false,"data":{"magic":2,"id":361037525172240,"tweetId":"361037525172240","gmtCreate":1729179441786,"gmtModify":1729179443408,"author":{"id":3489542956853100,"idStr":"3489542956853100","authorId":3489542956853100,"authorIdStr":"3489542956853100","name":"期权我就看看","avatar":"https://static.tigerbbs.com/51ccadd35dd38328efbbd9d4c7c9fb33","vip":1,"userType":1,"introduction":"","boolIsFan":false,"boolIsHead":false,"crmLevel":2,"crmLevelSwitch":0,"individualDisplayBadges":[],"fanSize":81,"starInvestorFlag":false},"themes":[],"images":[],"coverImages":[],"html":"<html><head></head><body><a href=\"https://laohu8.com/S/NVDA\">$英伟达(NVDA)$ </a>就一个问题，靠算力堆出来的人工智能，是真的人工智能吗，难道不是更应该靠神经网络算法的不断升级吗？</body></html>","htmlText":"<html><head></head><body><a href=\"https://laohu8.com/S/NVDA\">$英伟达(NVDA)$ </a>就一个问题，靠算力堆出来的人工智能，是真的人工智能吗，难道不是更应该靠神经网络算法的不断升级吗？</body></html>","text":"$英伟达(NVDA)$ 就一个问题，靠算力堆出来的人工智能，是真的人工智能吗，难道不是更应该靠神经网络算法的不断升级吗？","highlighted":1,"essential":1,"paper":1,"likeSize":4,"commentSize":7,"repostSize":0,"favoriteSize":0,"link":"https://laohu8.com/post/361037525172240","repostId":0,"isVote":1,"tweetType":1,"viewCount":3437,"commentLimit":10,"likeStatus":false,"favoriteStatus":false,"reportStatus":false,"symbols":["NVDA"],"verified":2,"subType":0,"readableState":1,"langContent":"CN","currentLanguage":"CN","warmUpFlag":false,"orderFlag":false,"shareable":true,"causeOfNotShareable":"","featuresForAnalytics":[],"commentAndTweetFlag":false,"andRepostAutoSelectedFlag":false,"upFlag":false,"length":108,"xxTargetLangEnum":"ZH_CN"},"commentList":[{"id":361516621889800,"commentId":"361516621889800","gmtCreate":1729279602368,"gmtModify":1729279606204,"authorId":4171124507330002,"author":{"id":4171124507330002,"idStr":"4171124507330002","authorId":4171124507330002,"name":"Patrick Tsui","avatar":"https://static.tigerbbs.com/ca1d142f16ba16f4c7c18e1967d44ada","vip":1,"crmLevel":2,"crmLevelSwitch":0,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":0,"supIdStr":"0","prevId":0,"prevIdStr":"0","content":"去Google算法和神经网路又不难，别只在网路上见过这两个词就觉得自己懂ai🤡","text":"去Google算法和神经网路又不难，别只在网路上见过这两个词就觉得自己懂ai🤡","html":"去Google算法和神经网路又不难，别只在网路上见过这两个词就觉得自己懂ai🤡","likeSize":0,"commentSize":1,"subComments":[{"id":379440012009896,"commentId":"379440012009896","gmtCreate":1733670874028,"gmtModify":1733707682109,"authorId":3489542956853100,"author":{"id":3489542956853100,"idStr":"3489542956853100","authorId":3489542956853100,"name":"期权我就看看","avatar":"https://static.tigerbbs.com/51ccadd35dd38328efbbd9d4c7c9fb33","vip":1,"crmLevel":2,"crmLevelSwitch":0,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":361516621889800,"supIdStr":"361516621889800","prevId":0,"prevIdStr":"0","content":"嗯，还是你牛逼，会用google搜算法理论就懂AI了","text":"嗯，还是你牛逼，会用google搜算法理论就懂AI了","html":"嗯，还是你牛逼，会用google搜算法理论就懂AI了","disclaimerType":0}],"verified":10,"allocateAmount":0,"commentType":"hot","coins":0,"score":0,"disclaimerType":0},{"id":361590613733680,"commentId":"361590613733680","gmtCreate":1729297722393,"gmtModify":1729297726103,"authorId":3553256311875125,"author":{"id":3553256311875125,"idStr":"3553256311875125","authorId":3553256311875125,"name":"老当益壮铁树开花","avatar":"https://static.tigerbbs.com/ffe9bdf2e928a90ce7307de1ff8b30ff","vip":1,"currentWearingBadge":{"badgeId":"35ec162348d5460f88c959321e554969-3","templateUuid":"35ec162348d5460f88c959321e554969","name":"传说交易员","description":"证券或期货账户累计交易次数达到300次","bigImgUrl":"https://static.tigerbbs.com/656db16598a0b8f21429e10d6c1cb033","smallImgUrl":"https://static.tigerbbs.com/03f10910d4dd9234f9b5702a3342193a","grayImgUrl":"https://static.tigerbbs.com/0c767e35268feb729d50d3fa9a386c5a","redirectLinkEnabled":0,"hasAllocated":1,"isWearing":1,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2021.12.21","exceedPercentage":"93.62%","individualDisplayEnabled":0,"individualDisplaySort":0,"categoryType":1100},"crmLevel":6,"crmLevelSwitch":0,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":0,"supIdStr":"0","prevId":0,"prevIdStr":"0","content":"就跟曾经的超级计算机一样，靠堆cpu实现的算力不及10年后的一颗cpu算力，其实都是资本的阴谋，否则哪来翻倍的想象力。当然很多所谓懂AI的人要嗤之以鼻了。","text":"就跟曾经的超级计算机一样，靠堆cpu实现的算力不及10年后的一颗cpu算力，其实都是资本的阴谋，否则哪来翻倍的想象力。当然很多所谓懂AI的人要嗤之以鼻了。","html":"就跟曾经的超级计算机一样，靠堆cpu实现的算力不及10年后的一颗cpu算力，其实都是资本的阴谋，否则哪来翻倍的想象力。当然很多所谓懂AI的人要嗤之以鼻了。","likeSize":0,"commentSize":1,"subComments":[{"id":379425919828416,"commentId":"379425919828416","gmtCreate":1733671021948,"gmtModify":1733671025646,"authorId":3489542956853100,"author":{"id":3489542956853100,"idStr":"3489542956853100","authorId":3489542956853100,"name":"期权我就看看","avatar":"https://static.tigerbbs.com/51ccadd35dd38328efbbd9d4c7c9fb33","vip":1,"crmLevel":2,"crmLevelSwitch":0,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":361590613733680,"supIdStr":"361590613733680","prevId":0,"prevIdStr":"0","content":"哈哈，看看你楼下那位不就是吗，就是因为算法还不行效率太低了所以才需要上大算力才能出结果，但是算力是马上花钱买卡能解决的，算法可不是一时半会儿就能有质的飞跃，多数人只能看到眼前的利益","text":"哈哈，看看你楼下那位不就是吗，就是因为算法还不行效率太低了所以才需要上大算力才能出结果，但是算力是马上花钱买卡能解决的，算法可不是一时半会儿就能有质的飞跃，多数人只能看到眼前的利益","html":"哈哈，看看你楼下那位不就是吗，就是因为算法还不行效率太低了所以才需要上大算力才能出结果，但是算力是马上花钱买卡能解决的，算法可不是一时半会儿就能有质的飞跃，多数人只能看到眼前的利益","disclaimerType":0}],"verified":10,"allocateAmount":0,"commentType":"hot","coins":0,"score":0,"disclaimerType":0},{"id":361191488348400,"commentId":"361191488348400","gmtCreate":1729217031325,"gmtModify":1729217035068,"authorId":3574720750802706,"author":{"id":3574720750802706,"idStr":"3574720750802706","authorId":3574720750802706,"name":"Xzhouz","avatar":"https://static.tigerbbs.com/8056d36a8d17266379632873b136f02c","vip":1,"currentWearingBadge":{"badgeId":"1f74612b6f95455f970edd02b5316db8-3","templateUuid":"1f74612b6f95455f970edd02b5316db8","name":"收益排行","description":"Top 20%","bigImgUrl":"https://static.tigerbbs.com/1ece4391160b811e6f36b1eea9eb6ea4","smallImgUrl":"https://static.tigerbbs.com/7d7540817554a583ac3bf85dea074a4a","grayImgUrl":"https://static.tigerbbs.com/8c5460a7cc80dd22560efa26813d947d","redirectLinkEnabled":0,"redirectLink":"https://www.itigerup.com/activity/market/2023/stock-contest?adcode=StockContest2023","hasAllocated":1,"isWearing":1,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2024.07.13","individualDisplayEnabled":0,"individualDisplaySort":0,"categoryType":5300},"hat":"https://static.tigerbbs.com/f3e7f6253c885de8bad6403055b76aab","crmLevel":2,"crmLevelSwitch":1,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":0,"supIdStr":"0","prevId":0,"prevIdStr":"0","content":"这也是我一直思考的问题，如果堆量可以实现，那么买三五个低端卡代替一个高端卡，能解决这个问题。剩下的无非是成本和能耗的问题。所以制约人工智能的最重要的因素到底是高端显卡，还是算力，还是软件？","text":"这也是我一直思考的问题，如果堆量可以实现，那么买三五个低端卡代替一个高端卡，能解决这个问题。剩下的无非是成本和能耗的问题。所以制约人工智能的最重要的因素到底是高端显卡，还是算力，还是软件？","html":"这也是我一直思考的问题，如果堆量可以实现，那么买三五个低端卡代替一个高端卡，能解决这个问题。剩下的无非是成本和能耗的问题。所以制约人工智能的最重要的因素到底是高端显卡，还是算力，还是软件？","likeSize":0,"commentSize":2,"subComments":[{"id":361332125642880,"commentId":"361332125642880","gmtCreate":1729234571020,"gmtModify":1729239137488,"authorId":3504260094030003,"author":{"id":3504260094030003,"idStr":"3504260094030003","authorId":3504260094030003,"name":"二狗大石","avatar":"https://static.laohu8.com/default-avatar.jpg","vip":1,"crmLevel":5,"crmLevelSwitch":1,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":361191488348400,"supIdStr":"361191488348400","prevId":0,"prevIdStr":"0","content":"不冲突啊 算力是基础 神经网络算法也是基础 两者都要更好才能更好 至于3个低端卡换一个高端卡 是完全不划算的 多个低端卡提供高端卡一样的算力 能耗是更高的","text":"不冲突啊 算力是基础 神经网络算法也是基础 两者都要更好才能更好 至于3个低端卡换一个高端卡 是完全不划算的 多个低端卡提供高端卡一样的算力 能耗是更高的","html":"不冲突啊 算力是基础 神经网络算法也是基础 两者都要更好才能更好 至于3个低端卡换一个高端卡 是完全不划算的 多个低端卡提供高端卡一样的算力 能耗是更高的","disclaimerType":0},{"id":361335024918632,"commentId":"361335024918632","gmtCreate":1729235191868,"gmtModify":1729235196255,"authorId":3504260094030003,"author":{"id":3504260094030003,"idStr":"3504260094030003","authorId":3504260094030003,"name":"二狗大石","avatar":"https://static.laohu8.com/default-avatar.jpg","vip":1,"crmLevel":5,"crmLevelSwitch":1,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":361037525172240,"objectIdStr":"361037525172240","type":1,"supId":361191488348400,"supIdStr":"361191488348400","prevId":0,"prevIdStr":"0","content":"我跑过一些简单的人工智能模型 例如图片的智能识别 第一 需要合适的算法以及正确的标注 第二 需要高端的显卡才能快速跑完一次学习 如果算力不够 跑一次训练就要很多天 那就效率太低了 如果我一天就能跑完一次 你要100天才能跑完一次 那效率肯定是天壤之别的（跑完以后得到的模型又要不断的测试优化 真实的开发是需要跑无数次）","text":"我跑过一些简单的人工智能模型 例如图片的智能识别 第一 需要合适的算法以及正确的标注 第二 需要高端的显卡才能快速跑完一次学习 如果算力不够 跑一次训练就要很多天 那就效率太低了 如果我一天就能跑完一次 你要100天才能跑完一次 那效率肯定是天壤之别的（跑完以后得到的模型又要不断的测试优化 真实的开发是需要跑无数次）","html":"我跑过一些简单的人工智能模型 例如图片的智能识别 第一 需要合适的算法以及正确的标注 第二 需要高端的显卡才能快速跑完一次学习 如果算力不够 跑一次训练就要很多天 那就效率太低了 如果我一天就能跑完一次 你要100天才能跑完一次 那效率肯定是天壤之别的（跑完以后得到的模型又要不断的测试优化 真实的开发是需要跑无数次）","disclaimerType":0}],"verified":10,"allocateAmount":0,"commentType":"hot","coins":0,"score":0,"disclaimerType":0}],"isCommentEnd":false,"isTiger":false,"isWeiXinMini":false,"url":"/m/post/361037525172240"}
精彩评论