av无码不卡在线观看免费,性行为视频在线观看无码

Terry Lambert
DeepSeek is an Open Source LLM (Large Language Model). ChatGPT is not Open Source, even though its value is not really in the code used to implement, it’s in the training datasets, and the curation choices made in the training. DeepSeek has a number of “strongly held beliefs”, if one can — as apparently, many people do — anthropomorphize a product of AI research, which is neither conscious, nor self aware, and actually holds no beliefs whatsoever.

DeepSeek 是一個開源的LLM（大型語言模型）。
ChatGPT 不是開源的，盡管它的價值并不在于用于實現(xiàn)的代碼，而在于訓(xùn)練數(shù)據(jù)集和訓(xùn)練中做出的管理選擇。
DeepSeek 有許多“堅定的信念”。人工智能是既沒有意識，也沒有自我意識的產(chǎn)品，實際上沒有任何信念。如果一個人想的話，他可以將人工智能研究的產(chǎn)品擬人化。顯然，很多人都是這么做的。

But if you have a conversation with DeepSeek, you will come away from it with the impression that it has a strong bias towards China’s world view, particularly if you ask it questions about religions which are banned in China. This is not a product of actual “strongly held beliefs”, it is a consequence of the data on which it operates in order to arrive at its answers, and the bias is inherent in the curation of that data. ChatGPT has a number of biases of its own, in that regard.

但如果你與 DeepSeek 進行對話，你會覺得它和中國的世界觀一樣有著強烈的偏見，特別是當你問它關(guān)于中國禁止的宗教的問題時。
這不是實際的“堅定的信仰”，而是為了特定的答案操作數(shù)據(jù)的結(jié)果，這種偏見是數(shù)據(jù)管理中固有的。
在這方面，ChatGPT 自己也存在著很多偏見。
原創(chuàng)翻譯：龍騰網(wǎng) http://nxnpts.cn 轉(zhuǎn)載請注明出處

The main thing about DeepSeek, however, is that you can train it with your own training data, curate it the way you want to curate it, and you will end up with a great diagnostician,or plumbing repair suggestions which can actually fix problems that other plumbers have run into, based on their life experience. Or you can create something that responds like a David Duke or Milo Yiannopoulos or Al Gore or Bernie Sanders sycophant, if you ask it political questions.

然而，DeepSeek 最主要的一點是，你可以用你自己的訓(xùn)練數(shù)據(jù)對它進行訓(xùn)練，按照你想要的方式進行管理，最終你會得到一個很棒的診斷專家。
比如：根據(jù)其他水管工的生活經(jīng)驗，提出可以真正解決他們遇到的問題的管道維修建議。
或者，如果你問它政治問題，你可以創(chuàng)造一個像 David Duke 或 Milo Yiannopoulos 或 Al Gore 或 Bernie Sanders 那樣的馬屁精一樣的東西。

The point being is that the bias is largely in the training data, and that, as a tool, Open Source is a better mechanism for distribution and code improvement over time, especially given that it’s the curated training set data that is strategic, and the code that operates on that data,is merely tactical. I’ve suggested to ChatGPT, and to the scientists at ChatGPT whom I know personally, that it’s a mistake to keep it closed source. ChatGPT should be Open Sourced to protect the leadership inherent in its tactical code model. But I have no doubt that DeepSeek, if run against the ChatGPT 4 training set data.would almost certainly come up with answers indistinguishable, for the most part, from the ChatGPT 4 answers you’d get, with only minor variations.

關(guān)鍵在于，偏見主要存在于訓(xùn)練數(shù)據(jù)中。
作為一種工具，開源是一種更好的分發(fā)和代碼改進機制，尤其是考慮到精心策劃的訓(xùn)練集數(shù)據(jù)是戰(zhàn)略性的，而對這些數(shù)據(jù)進行操作的代碼僅僅是戰(zhàn)術(shù)性的。
我曾向 ChatGPT 以及我認識的 ChatGPT 科學(xué)家建議，保持閉源是一個錯誤。
ChatGPT 應(yīng)該開源，以保護其戰(zhàn)術(shù)代碼模型固有的領(lǐng)導(dǎo)地位。
但我毫不懷疑，如果用針對 ChatGPT 4 的訓(xùn)練集數(shù)據(jù)運行 DeepSeek，它得到的答案幾乎肯定會與你在ChatGPT 4得到的答案差不多一致，只有細微的差別。

Now ask China to share its DeepSeek training data, run your own DeepSeek (if you can afford the hardware), and see if it gives you the same, or different answers, from DeepSeek itself. I suspect any differences will be because they did not share some of their training data, because of how that training data would look to the rest of the world. But at least the code itself, tactical and not strategic though it is, is Open Source.

現(xiàn)在要求中國分享其 DeepSeek 的訓(xùn)練數(shù)據(jù)，運行您自己的 DeepSeek（如果您能負擔得起硬件的話），你可以看看，他們給出的答案是否會想同。
我懷疑任何差異，都是因為他們沒有分享他們的一些訓(xùn)練數(shù)據(jù)，因為這些訓(xùn)練數(shù)據(jù)在世界其他地方看起來就是那個樣子的。
但至少代碼本身是開源的，盡管它是戰(zhàn)術(shù)性的而不是戰(zhàn)略性的。

Paul Denlinger You can download the source code for DeepSeek from GitHub and you can modify it locally running on your own hardware. If you want to modify your system to hate the Chinese government, you can do it. On the other hand, I cannot stop ChatGPT from censoring its own answers when it comes to the Israel/Palestine issue because it is a proprietary system and no one can access the source code.

你可以從 GitHub 下載 DeepSeek 的源代碼，然后在自己本地的硬件上修改它。如果你想把它修改成憎恨中國政府的系統(tǒng)，你也可以這么做。
另一方面，我也無法阻止 ChatGPT 在涉及以色列/巴勒斯坦問題時審查它自己的答案，因為它是一個專有系統(tǒng)，沒有人可以訪問源代碼。

Russell Miller
Is there such a thing as unbiased AI?

是否存在無偏見的人工智能？

Paul Denlinger
No, that is why you should download DeepSeek and make your own local version of DeepSeek which gives you the answers you want.

沒有，這就是為什么你應(yīng)該下載 DeepSeek，并制作你自己的本地版本的 DeepSeek ，來給你你想要的答案的原因。

Sidney Kan
In your sense, yes. You can make it unbiased in your sense. In practical sense, no. There is nothing absolutely unbiased. Unbiased is a theoretical concept that cannot be achieved by any realization.

從你的意義上來說，有的。你可以從你的意義上讓它不偏不倚。從實際意義上來說，沒有。沒有絕對不偏不倚的東西。不偏不倚是一個理論概念，任何現(xiàn)實都無法實現(xiàn)。

Luc De Mestre
The model is open source. You can fine-tune it for a very low price. If you want a DeepSeek with your exact political opinions, you can have it. But in reality, the main usage of LLMs is not to generate answers on Quora or to seek information. LLMs have a basic understanding of human language, and that is insanely valuable for millions of jobs. Hundred of millions of people work is basically reading an e-mail or discussing with somebody, then filling an excel sheet or a form out of the content of the mail. And that used to be 100% manual, because computers sucked at reading natural language.

該模型是開源的。你可以以非常低的價格對其進行微調(diào)。
如果你想要一個具有你確切政治觀點的 DeepSeek，你可以擁有它。
但實際上，LLM（大型語言模型）的主要用途不是在 Quora 上生成答案或?qū)ふ屹Y料。
LLM（大型語言模型）對人類語言有基本的了解，這對數(shù)百萬種工作來說非常有價值。數(shù)億人的工作基本上是閱讀電子郵件或與某人討論，然后根據(jù)郵件內(nèi)容填寫 excel 表或表格。而這曾經(jīng)是 100% 手動的，因為計算機在閱讀自然語言方面很糟糕。

A simple phrase like “Brian is making a sandwich in the kitchen” followed by the question “Where is Brian?” used to be an unsolved problem until the emergence of LLMs (basically, before 2017 that was impossible for computers). When you talk to your bank, to your insurance, to a customer service, they have a simple decision tree. The fact that you are covered or not by your insurance, that you are worthy or not of a loan, that you will get a refund or not, does not depends on a human being. Workers are trained with a strict set of rules.

一個簡單的短句，例如“布萊恩正在廚房做三明治”，后面跟著一個問題“布萊恩在哪里？”，在LLM 出現(xiàn)之前，這曾經(jīng)是一個未能解決的問題，基本上，在 2017 年之前，計算機不可能解決這個問題。
當你與銀行、保險公司、客服交談時，他們會有一個簡單的決策樹。你是否在保險范圍內(nèi)，你是否值得貸款，你是否能得到退款，這些都不取決于人。工作人員接受了一套嚴格的規(guī)則培訓(xùn)。

The work of the human you interact with is to translate your mail, or a conversation with you, into a few datapoints. You will tell a little story, and from that story a few numbers and facts will be extracted. And only human were able to do that in the past. “So you missed your plane and wants your ticket to be refunded. But why did you missed it? Because your dog was sick? Sorry, there is no refund. But if it was your mother, you would have been refunded.” And now LLMs can do that. They sucks at answering at Quora, but nobody is making money answering question on Quora. They sucks at writing novels, but nobody is making money writing novels (that is not a joke, most writers have a side-job or are super-poor).

與您互動的人的工作，是將您的郵件或與您的對話，轉(zhuǎn)換為幾個數(shù)據(jù)點。
您將講述一個小故事，他們會從該故事中提取一些數(shù)字和事實。過去只有人類才能做到這一點。
“所以你錯過了飛機，想退票。但你為什么錯過了？因為你的狗生病了？對不起，不能退款。但如果是你的母親生病了，你就可以被退款?！?br /> 現(xiàn)在LLM可以做到這一點。
他們在 Quora 上回答問題的時候很糟糕，但沒有人通過回答 Quora 上的問題賺錢。
他們在寫小說方面很糟糕，但沒有人通過寫小說賺錢（這不是開玩笑，大多數(shù)作家都有兼職或非常貧窮）。

But millions of people are making money at understanding basic English then filling a form summarizing an e-mail or a conversation. Each interaction with your bank, or your insurance, or any administration, is basically that. Because procedures are everything, and most workers have no decisional power at all. And DeepSeek is better at that task than any other LLM, with the exception of OpenAI O1, but OpenAI O1 is very very expensive. DeepSeek is faster, cheaper and can be run locally.

但數(shù)以百萬計的人通過理解基本英語，然后填寫總結(jié)電子郵件或?qū)υ挼谋砀駚碣嶅X。
與銀行、保險公司或任何行政部門的每次互動基本上都是這樣。
因為程序就是一切，而大多數(shù)打工人根本就沒有決策權(quán)。
DeepSeek 比其他任何 LLM 都更擅長這項任務(wù)，OpenAI O1 除外，但 OpenAI O1 非常非常昂貴。
DeepSeek 更快、更便宜，并且可以在本地運行。

And for reading a simple mail and deciding filling a form out of it, political bias are irrelevant. You will never ask a refund to an airplane company because of a xxx-square-related problem.

對于閱讀一封簡單的郵件，并決定如何填寫表格而言，政治偏見是無關(guān)緊要的。
你永遠不會因為xx門廣場相關(guān)的問題而向飛機公司要求退款。

Flavio Massimo Foglia
Probably because not all people want to use ChatGPT AI when its answers are so obviously biased toward the American worldview propaganda

可能是因為 ChatGPT AI 的答案，明顯偏向美國宣傳的世界觀，所以，并不是所有人都愿意使用它。

Sangram Sagar
This is an interesting and complex question. There’s no doubt that AI systems, including DeepSeek, can be influenced by the data they’re trained on, and that can result in bias—whether it’s political, cultural, or otherwise. When it comes to DeepSeek, some critics argue that its answers may reflect a Chinese-centric viewpoint, which could seem biased, especially when compared to other systems that might have more diverse or Western influences. But there are a few reasons why people are still adopting it, despite these concerns.

這是一個有趣而復(fù)雜的問題。
毫無疑問，包括 DeepSeek 在內(nèi)的人工智能系統(tǒng)會受到其訓(xùn)練數(shù)據(jù)的影響，并導(dǎo)致偏見——無論是政治、文化還是其他方面。
談到 DeepSeek，一些批評者認為，它的答案可能反映了以中國為中心的觀點，這似乎是偏見，尤其是與其他可能具有更多多樣性或西方影響的系統(tǒng)相比時。
但盡管存在這些擔憂，人們?nèi)匀徊捎盟脑蛴袔讉€。

First, there’s the issue of AI's usefulness in solving problems. Many users find that DeepSeek’s AI performs exceptionally well in terms of data analysis, problem-solving, and specialized tasks, especially in industries like technology, manufacturing, and engineering. For users in China or regions closely aligned with it, this can be incredibly valuable. In these cases, the utility of the tool outweighs any perceived bias in the responses.

首先，存在人工智能在解決問題方面的實用性問題。
許多用戶發(fā)現(xiàn) DeepSeek 的人工智能在數(shù)據(jù)分析、解決問題和專業(yè)任務(wù)方面表現(xiàn)非常出色，尤其是在技術(shù)、制造和工程等行業(yè)。
對于中國或與其密切相關(guān)的地區(qū)的用戶來說，這可能非常有價值。在這些情況下，該工具的實用性超過了響應(yīng)中任何可察覺的偏見。

Second, AI systems like DeepSeek are often designed with a specific user base in mind. In this case, China has a massive interest in developing AI technologies to help propel their global competitiveness. For domestic users, there’s less concern about propaganda since the content and solutions are tailored to their needs. When seen from this perspective, the potential bias in the system might be less of an issue compared to the benefits it provides in efficiency and innovation.

其次，像 DeepSeek 這樣的人工智能系統(tǒng)通常是為特定的用戶群設(shè)計的。
在這種情況下，中國對開發(fā)人工智能技術(shù)有著濃厚的興趣，以幫助提升其全球競爭力。
對于中國國內(nèi)用戶來說，他們不必太擔心宣傳，因為內(nèi)容和解決方案都是根據(jù)他們的需求量身定制的。
從這個角度來看，與它在效率和創(chuàng)新方面帶來的好處相比，該系統(tǒng)中潛在的偏見可能不是什么大問題。

Moreover, it's important to understand that AI bias isn't unique to DeepSeek. Many other AI systems—including those developed in the West—can have biases based on the data they’re trained on or the cultural contexts they stem from. The difference is that with DeepSeek, because it’s deeply integrated into China’s technology ecosystem, users there may be more likely to embrace it because it aligns with their values, even if some might perceive it as biased.

此外，我們必須明白，人工智能偏見并非 DeepSeek 獨有。
許多其他人工智能系統(tǒng)（包括在西方開發(fā)的系統(tǒng)）都可能存在基于其訓(xùn)練數(shù)據(jù)或文化背景的偏見。
不同之處在于，由于 DeepSeek 深度融入了中國的技術(shù)生態(tài)系統(tǒng)，中國用戶可能更有可能接受它，因為它符合他們的價值觀，即使有些人可能認為它有偏見。

Finally, the reality is that AI technology is becoming increasingly integrated into daily life globally. Some people may not fully understand or pay attention to the nuances of bias in these systems, focusing instead on their practical applications. DeepSeek offers an edge in certain tasks and industries, which keeps it appealing even for those who are concerned about its political slant. In short, the adoption of DeepSeek’s AI comes down to its practical effectiveness and how much users are willing to overlook the potential bias. People are drawn to useful technology, and if DeepSeek offers a significant advantage, they might be more inclined to adopt it—bias or not.

最后，現(xiàn)實情況是，人工智能技術(shù)正日益融入全球的日常生活。
有些人可能無法完全理解或注意這些系統(tǒng)中偏見的細微差別，而是專注于它們的實際應(yīng)用。
DeepSeek 在某些任務(wù)和行業(yè)中具有優(yōu)勢，這使得它即使對那些擔心其政治傾向的人也具有吸引力。
簡而言之，DeepSeek 人工智能的采用取決于其實際有效性，以及用戶愿意在多大程度上忽略其潛在的偏見。
人們被有用的技術(shù)所吸引，如果 DeepSeek 提供了顯著的優(yōu)勢，他們可能更傾向于采用它——無論是否具有偏見。

Aaron Kaufman
It’s not that they’re biased, but that they’re so blatantly wrong that bothers me..:

并不是說他們有偏見，而是他們錯誤得如此明顯，令我困擾……

Stephen Ford The questioner is misinformed. DeepSeek is open source and can be trained using any data the users choose, so any bias they want, or no bias at all. The significance of DeepSeek is the reported huge reduction in scale to achieve what other models have done. If true this represents a vast quantum leap in efficiency. The lead DeepSeek currently has won’t last long because the techniques will be copied by everybody else. To prattle on about bias is just sheer ignorance about DeepSeek’s significance, which is about AI technology, not how it is applied.

提問者被誤導(dǎo)了。
DeepSeek 是開源的，可以使用用戶選擇的任何數(shù)據(jù)進行訓(xùn)練，因此可以使用他們想要的帶有任何偏見，或者根本不帶任何偏見的數(shù)據(jù)。
DeepSeek 的意義在于，據(jù)報道，它大大減少了規(guī)模，卻實現(xiàn)了其他模型所能做的事情。如果屬實，這代表著效率的巨大飛躍。
DeepSeek 目前所擁有的領(lǐng)先地位不會持續(xù)太久，因為這些技術(shù)將被其他人復(fù)制。
喋喋不休地談?wù)撈娭皇菍?DeepSeek 意義的無知，它關(guān)乎人工智能技術(shù)，而不是它如何應(yīng)用。
原創(chuàng)翻譯：龍騰網(wǎng) http://nxnpts.cn 轉(zhuǎn)載請注明出處

Dave Ch
BecauseDS is Chinese and follows Chinese rules and pov. You don't go to a Chinese restaurant and complain when they don't give a good cheeseburger. Here's some propoganda for you: Chinese people everywhere are smart, hardworking and results focused.

因為 DeepSeek 是中國的，所以遵守中國的規(guī)則和觀點。你不應(yīng)該去一家中餐館，然后抱怨他們沒有提供好吃的芝士漢堡。
我要對你提供你一點宣傳：世界各地的中國人都很聰明、勤奮、注重實用性。

Andreas Ioannou
People are adopting DeepSeek's AI because it offers a quick and convenient way to access information. While it may have biases towards the Chinese worldview propaganda, some people may still find value in the insights and perspectives it provides. Additionally, not everyone may be aware of the biases or may choose to overlook them in favor of the convenience and efficiency that the AI offers.

人們之所以采用 DeepSeek 的人工智能，是因為它提供了一種快速便捷的信息獲取方式。
盡管它可能偏向中國宣傳的世界觀，但有些人可能仍然認為它提供的見解和觀點很有價值。
此外，并不是每個人都意識到了這些偏見，他們可能選擇忽視它們，而選擇人工智能可以提供的便利和效率。

Freddie Chen
Because that AI is not a tool for biased people like you who think if the answer did not match his/her expectations, it is the biased propganda. Like Zionism think if you did not support it, you were anti Semitic. AI are helping people in a lot of industries that will not differ Chinese or US but giving an answer applies to obxtive fact. For exmple, 1+1 equals to 2. It is just a waste of you to use AI but not for other people.

因為人工智能是這樣一個工具，它不像你一樣，具有偏見。你們這樣的人，如果認為答案不符合自己的期望，那就是有偏見的宣傳。
就像猶太復(fù)國主義者認為如果你不支持它，你就是反猶太主義者。
人工智能正在幫助很多行業(yè)的人，這些行業(yè)不區(qū)分中國或美國，但給出答案適用于客觀事實。例如，1+1 等于 2。
你這樣的人使用人工智能就是浪費，而其他人不是。

David Wrixon
Because since Jan 6, we are all pivoting towards China. Still better than listening to Diaper Donnie.

因為從 1 月 6 日起，我們都轉(zhuǎn)向了中國。這比聽尿布唐（川普）的演講好多了。
譯注： Diaper Donnie ，尿布唐，川普的綽號，基于某次訪談，說他吸食某種物質(zhì)后（你懂的），經(jīng)常會大小便失禁，需要急忙去換褲子。真假未知。

David Horton
I'm using it now, seems ok to me I like its interface, simple ! It's not superior to ChatGPT … Probably inferior in many cases Time will tell …

我現(xiàn)在正在使用它，我覺得還不錯
我喜歡它的界面，簡單！
它并不比 ChatGPT 好……在很多情況下可能不如它。
時間會證明一切……

QA問答：為什么人們會采用DeepSeek的人工智能，它的答案明顯偏向于中國宣傳的世界觀？

QA問答：為什么人們會采用DeepSeek的人工智能，它的答案明顯偏向于中國宣傳的世界觀？