【Hacker News搬运】OpenAI与博通和台积电共同打造首款芯片，缩减代工野心

hackernews

Title: OpenAI builds first chip with Broadcom and TSMC, scales back foundry ambition

OpenAI与博通和台积电共同打造首款芯片，缩减代工野心

Text:

Url: https://www.reuters.com/technology/artificial-intelligence/openai-builds-first-chip-with-broadcom-tsmc-scales-back-foundry-ambition-2024-10-29/

很抱歉，作为一个AI，我无法直接访问外部链接或使用特定工具如JinaReader来抓取和分析网页内容。不过，我可以根据你提供的链接标题和我的知识来为你提供一些信息。

标题为 "OpenAI builds first chip with Broadcom, TSMC scales back foundry ambition - Reuters"，这篇Reuters的文章可能讨论了以下内容：

1. **OpenAI 与 Broadcom 合作**：OpenAI 与半导体公司 Broadcom 合作，共同开发了一款芯片。这可能是为了支持OpenAI在人工智能计算方面的需求，尤其是在处理大规模机器学习模型时。

2. **台积电缩减晶圆代工雄心**：台积电（TSMC）是全球最大的晶圆代工厂，文章可能提到TSMC在晶圆代工领域的战略调整。这可能包括减少某些业务的投资或关注特定领域的增长。

以下是对上述内容的中文总结：

“OpenAI与半导体巨头Broadcom合作，共同打造了其首款芯片。此举可能是为了满足OpenAI在人工智能计算方面的需求。同时，全球最大的晶圆代工厂台积电（TSMC）正在缩减其在晶圆代工领域的雄心，可能意味着其战略将更加专注于特定领域的发展。”

请注意，这只是一个基于标题的总结，具体内容可能有所不同。如果你需要更详细的信息，建议直接阅读原文。

Post by: marban

Comments:

jsheard: > OpenAI considered building everything in-house and raising capital for an expensive plan to build a network of factories known as "foundries" for chip manufacturing. The company has dropped the ambitious foundry plans for now due to the costs and time needed to build a networkThat framing massively undersells how insane Sams ambitions were there, he was floating the idea of somehow raising seven trillion dollars to build thirty six fabs dedicated to making AI silicon. The TSMC execs reported more or less laughed in his face when he brought it up it them.

jsheard: &gt；OpenAI考虑在内部建造一切，并为一项耗资巨大的计划筹集资金，以建立一个名为&quot；铸造厂”；由于构建网络所需的成本和时间，该公司目前已经放弃了雄心勃勃的代工计划。鉴于萨姆的野心有多么疯狂，他提出了以某种方式筹集7万亿美元来建造36家专门用于制造人工智能硅的晶圆厂的想法。台积电的高管们报告说，当他提起这件事时，他们或多或少都笑了。

kccqzy: Not surprising at all. As a former Google employee, I'm regularly in contact with friends still at Google and the number of hardware engineers OpenAI has hired out of Google have drastically increased in the last few months. You can readily see for yourself just by browsing LinkedIn and looking at profiles of hardware engineers currently at OpenAI and noticing how many of them are former Google employees. Google's own TPU team is now a shell of its former self. I wouldn't be surprised if future Google TPUs are inferior to the new OpenAI chips.Also, did you know that Google's TPU efforts previously also relied on help from Broadcom? <a href="https://www.theregister.com/AMP/2023/09/22/google_broadcom_tpus/" rel="nofollow">https://www.theregister.com/AMP/2023/09/22/google_broadcom_t...</a> It's absolutely unsurprising that former Google employees would bring this vendor relationship into their new company, given Broadcom did a good job helping Google with TPUs.

kccqzy: 一点也不奇怪。作为一名前谷歌员工，我；我经常与仍在谷歌的朋友联系，过去几个月，OpenAI从谷歌聘请的硬件工程师数量急剧增加。你只需浏览领英，查看目前在OpenAI的硬件工程师的个人资料，并注意到他们中有多少是前谷歌员工，就可以很容易地看到自己的情况。谷歌；中国自己的TPU团队现在已经是一个空壳了。我会；如果未来的谷歌TPU不如新的OpenAI芯片，也不要感到惊讶 此外，你知道谷歌；TPU的努力以前也依赖于博通的帮助吗<a href=“https://www.theregister.com”2023年9月22日AMP“google_broadcom_tpus”rel=“nofollow”>https:&#x2F；www.theregister.com；AMP；2023年；09；22°；google_broadcom_t…</a>它；考虑到博通在TPU方面帮助谷歌做得很好，前谷歌员工将这种供应商关系带入他们的新公司，这绝对不足为奇。

latchkey: <pre><code> "while adding AMD (AMD.O), opens new tab chips alongside Nvidia (NVDA.O), opens new tab chips to meet its surging infrastructure demands"
</code></pre>
Jensen has been saying that demand is "insane" and we're hearing rumors of low yields. This equates to supply issues in the coming months/years. No fortune 500 puts all their eggs into one basket. Diversifying away from a single source for all of AI hardware and software, is a smart thing to do.

latchkey: <前><代码>“；在加入AMD（AMD.O）的同时，与Nvidia（NVDA.O）一起推出了新的标签芯片，推出了新标签芯片以满足其激增的基础设施需求；</code></pre>Jensen一直说，需求是&quot；疯狂&quot；我们；再次听到低收益的传言。这相当于未来几个月的供应问题；年。没有财富500强把所有的鸡蛋放在一个篮子里。将所有人工智能硬件和软件从单一来源多样化是一件明智的事情。

HarHarVeryFunny: So presumably they're building their own data centers to use these chips, but odd that the article doesn't actually say so.Are any of the western AI/matmul chips NOT being fabbed by TSMC? Amazon or Meta's perhaps? We've got NVIDIA, AMD, Google (TPU - also designed by Broadcom), and now OpenAI all using them. Are Samsung not competitive?I wonder to what extent total chip/chipset volume is limited by TSMC capacity as opposed to memory or packaging?

HarHarVeryFunny: 因此，据推测他们；重建自己的数据中心以使用这些芯片，但奇怪的是，这篇文章没有；实际上我没有这么说。西方AI；matmul芯片不是由台积电制造的吗？亚马逊或Meta；也许吧？我们；我们有NVIDIA、AMD、谷歌（TPU，也由博通设计），现在OpenAI都在使用它们。三星没有竞争力吗 我想知道总芯片在多大程度上；芯片组的体积受到台积电容量的限制，而不是内存或封装？

AdieuToLogic: Considering the cogent financial analysis detailed in "The Subprime AI Crisis"[0], all I can add is a quote from the movie "Dodgeball"[1]:<pre><code> It's a bold strategy, Cotton. Let's see if it pays off for 'em.
</code></pre>
0 - <a href="https://www.wheresyoured.at/subprimeai/" rel="nofollow">https://www.wheresyoured.at/subprimeai/</a>1 - <a href="https://www.imdb.com/title/tt0364725/quotes/" rel="nofollow">https://www.imdb.com/title/tt0364725/quotes/</a>

AdieuToLogic: 考虑到&quot；次贷AI危机&quot；[0]，我只能补充一句电影中的一句话&quot；道奇球&quot；[1] ：<pre><code>它；这是一个大胆的策略，科顿。让；让我们看看这是否值得；相对长度单位。</code></pre>0-<a href=“https:&#x2F；www.wheresyoured.at#x2F subprimeai#x2F”rel=“nofollow”>https:&#x2F；www.whethersyoured.at&#x2F；次级优先级</a> 1-<a href=“https:”www.imdb.com“title”tt0364725“quotes”rel=“nofollow”>https:”&#x2F；www.imdb.com；title；tt0364725；引号&#x2F</一