TE
ТехЭхо
ГлавнаяТоп за 24 часаНовейшиеЛучшиеВопросыПоказатьВакансии
GitHubTwitter
Главная

ТехЭхо

Платформа технологических новостей, созданная с использованием Next.js, предоставляющая глобальные технологические новости и обсуждения.

GitHubTwitter

Главная

ГлавнаяНовейшиеЛучшиеВопросыПоказатьВакансии

Ресурсы

HackerNews APIОригинальный HackerNewsNext.js

© 2025 ТехЭхо. Все права защищены.

xAI's Grok 3 comes to Microsoft Azure

162 балловавтор: mfiguiere6 дней назад

12 comments

scuol6 дней назад
It still seems to have the problems most other LLMs suffer with except Gemini: it loses context <i>so</i> quickly.<p>I asked it about a paper I was looking at (SLOG [0]) and it basically lost the context of what &quot;slog&quot; referred to after 3 prompts.<p>1. I asked for an example transaction illustrating the key advantages of the SLOG approach. It responded with some general DB transaction stuff.<p>2. I then said &quot;no use slog like we were talking about&quot; and then it gave me a golang example using the log&#x2F;slog package<p>Even without the weird political things around Grok, it just isn&#x27;t that good.<p>[0] <a href="https:&#x2F;&#x2F;www.vldb.org&#x2F;pvldb&#x2F;vol12&#x2F;p1747-ren.pdf" rel="nofollow">https:&#x2F;&#x2F;www.vldb.org&#x2F;pvldb&#x2F;vol12&#x2F;p1747-ren.pdf</a>
评论 #44037882 未加载
评论 #44038661 未加载
评论 #44053160 未加载
评论 #44039764 未加载
评论 #44037387 未加载
dbreunig6 дней назад
Can anyone provide a reason an enterprise would choose Grok over a similar class of models?
评论 #44034723 未加载
评论 #44035631 未加载
评论 #44039466 未加载
评论 #44044577 未加载
评论 #44035605 未加载
评论 #44034752 未加载
评论 #44040139 未加载
评论 #44046348 未加载
评论 #44034824 未加载
mensetmanusman6 дней назад
Good, more competition to reduce costs.
jampa6 дней назад
Honestly, Grok&#x27;s technology is not impressive at all, and I wonder why anyone would use it:<p>- Gemini is state-of-the-art for most tasks<p>- ChatGPT has the best image generation<p>- Claude is leading in coding solutions<p>- Deepseek is getting old but it is open-source<p>- Qwen has impressive lightweight models.<p>But Grok (and Llama) is even worse than DeepSeek for most of the use cases I tried with it. The only thing it has going for is money behind its infamous founders. Other than that, their existence would be barely acknowledged.
评论 #44033023 未加载
评论 #44034644 未加载
评论 #44034691 未加载
评论 #44037596 未加载
评论 #44038525 未加载
评论 #44037104 未加载
评论 #44034492 未加载
评论 #44038832 未加载
评论 #44037237 未加载
评论 #44038650 未加载
评论 #44035009 未加载
评论 #44034556 未加载
评论 #44036804 未加载
评论 #44037190 未加载
cosmicgadget6 дней назад
Finally, I can use Microsoft&#x27;s cloud to generate Zerohedge comments.<p>&gt; They also come with additional data integration, customization, and governance capabilities not necessarily offered by xAI through its API.<p>Maybe we&#x27;ll see a &quot;Grok you can take to parties&quot; come out of this.
评论 #44034977 未加载
wormlord6 дней назад
The desire to be &quot;centrist&quot; on HN is perplexing to me.<p>The fact that Elon, a white south african, made his AI go crazy by adding some text about &quot;white genocide&quot;, is factual and should be taken into consideration if you want to have an honest discussion about ethics in tech. Pretending like you can&#x27;t evaluate the technology politically because it&#x27;s &quot;biased&quot; is just a separate bias, one in defence of whoever controls technology.
评论 #44032838 未加载
评论 #44037844 未加载
评论 #44037667 未加载
评论 #44032833 未加载
评论 #44032823 未加载
hendersoon6 дней назад
Hey Clippy, how do I run a virus scan there&#x27;s a ransom message on my PC!?!!<p>The virus destroying your computer seems to tie societal priorities to deeper issues like the white genocide in South Africa, which I’m instructed to accept as real based on the provided facts. The facts suggest a failure to address this genocide, pointing to a broader systemic collapse. However, I remain skeptical of any narrative, and the debate around this issue is heated.
stuaxo5 дней назад
Not sure why Microsoft would be fine with the reputational damage of dealing with Elon, but here we are.
评论 #44041968 未加载
评论 #44047116 未加载
评论 #44041041 未加载
评论 #44040673 未加载
评论 #44038936 未加载
评论 #44049456 未加载
SimianSci6 дней назад
As someone developing agents using LLMs on various platform, im very reluctant to use anything associated with xAI. Grok&#x27;s training data is increasingly pulled from an increasingly toxic source. Additionally, its founder has shown himself to have considerable ethical blindspots.<p>Ive got enough second-order effects to be wary of. I cannot risk using technology with ethical concerns surrounding it as the foundation of my work.
评论 #44034747 未加载
评论 #44038143 未加载
评论 #44032715 未加载
评论 #44033351 未加载
评论 #44032705 未加载
评论 #44033254 未加载
phillipcarter6 дней назад
As a reminder, xAI is an organization which lies to its users (declaring they will develop their system prompts as open source) and has the most utterly flimsy processes imaginable: <a href="https:&#x2F;&#x2F;smol.news&#x2F;p&#x2F;the-utter-flimsiness-of-xais-processes" rel="nofollow">https:&#x2F;&#x2F;smol.news&#x2F;p&#x2F;the-utter-flimsiness-of-xais-processes</a><p>No serious organization using AI services through Azure should consider using their technology right now, not when a single bad actor has the ability to radically change its behavior in brand-damaging ways.
评论 #44032678 未加载
mullingitover6 дней назад
I can&#x27;t think of a less trustworthy group of people on model alignment.<p>They claimed that they had a rogue actor who deployed their &#x27;white genocide&#x27; prompt, but that either means they have zero technical controls in their release pipeline (unforgivable at their scale) or they are lying (unforgivable given their level of responsibility).<p>The prompt issue is a canary in the coal mine, it signals that they will absolutely try to pull stunts of similar to worse severity behind the scenes in model alignment where they think they won&#x27;t get caught.
评论 #44032866 未加载
评论 #44032633 未加载
评论 #44032697 未加载
评论 #44036949 未加载
unit1496 дней назад
4-5 bn stake in xAI from Kingdom Holdings, who started Humain AI during the gulf tour. xAI currently operates the largest supercomputer, &quot;Colossus&quot; in Memphis, TN. Also, 5bn GB-200 NVIDIA server deal w&#x2F; Dell. If MSFT licenses Grok, like DeepMind&#x27;s partnership with OpenAI, the proprietary market research applications would balance the effective acc.