chatgpt 会改变世界吗?

Posted on Sun 05 March 2023 in Journal

Abstract Journal on 2023-03-05
Authors Walter Fan
 Category    learning note  
Status v1.0
Updated 2023-03-05
License CC-BY-NC-ND 4.0

ChatGPT 是最近大火的 AI 应用, 让无数英雄竞折腰. 据说它还能干不少咱们程序员的活.

它是一个基于GPT-3(Generative Pretrained Transformer 3)的语言模型,用于处理自然语言问答, 是由人工智能公司OpenAI开发的一种大型神经网络模型,能够处理自然语言文本。ChatGPT是基于GPT-3模型构建的,能够根据用户输入的问题,生成自然语言的回答。

我用 https://sms-activate.org/ 申请了个帐号, 用来接收短信, 选一个英国的虚拟号码, 通过支付宝付了一美元就能搞定, chatgpt 的帐号申请也挺简单, 输入邮件, 手机号码就搞定了.

测试一

我问了它一个问题

Q. 如何用 python 画一个拆线图和相应的数据表格?

这是它给我的回答:

要用 Python 绘制拆线图和数据表格,可以使用 Matplotlib 和 Pandas 库。

以下是一个简单的示例代码,可以帮助您画出一张拆线图和相应的数据表格:

import matplotlib.pyplot as plt
import pandas as pd

# 构造数据
data = {'year': [2010, 2011, 2012, 2013, 2014, 2015, 2016],
        'sales': [12, 18, 23, 15, 17, 21, 19]}
df = pd.DataFrame(data)

# 绘制拆线图
plt.plot(df['year'], df['sales'], marker='o')
plt.xlabel('Year')
plt.ylabel('Sales')
plt.title('Sales Trend')

# 绘制数据表格
fig, ax = plt.subplots()
ax.axis('off')
ax.axis('tight')
ax.table(cellText=df.values, colLabels=df.columns, loc='center')

# 显示图表
plt.show()

在上面的代码中,我们首先构造了一个包含年份和销售额的数据框,然后使用 Matplotlib 绘制了一张拆线图,并为图表添加了适当的标签和标题。接着,我们创建了一个新的图表,绘制了数据表格,并将其置于中心位置。

运行代码后,会弹出一个窗口,其中包含了我们绘制的拆线图和数据表格。您也可以将图表保存为图像文件,使用 plt.savefig() 方法即可。

注意,为了使数据表格能够正确地显示,我们需要将其置于一个新的图表中,并将图表的坐标轴关闭。此外,还需要将数据框转换为一个包含字符串值的二维数组,以便将其传递给 ax.table() 方法。

评价

这个回答怎么样呢? 我觉得回答得挺好, 起码如果有人问我这个问题,我回答不了这么好, 顶多讲一下用什么库, 再给个自己写的例子程序.

测试二

这只是牛刀小试, 我以问了它另外一个问题

Q. what's the difference between REMB and TWCC in WebRTC congestion control ?

REMB (Receiver Estimated Maximum Bitrate) and TWCC (Transport Wide Congestion Control) are both techniques used in WebRTC congestion control, but they serve different purposes.

REMB is a protocol used by the receiver to estimate the maximum bitrate it can receive from the sender in real-time communication. It calculates the maximum bitrate based on the amount of bandwidth available in the network and the receiver's ability to handle the incoming data. This helps the sender to optimize the transmission of audio and video data by adjusting the bitrate of the data stream to avoid congestion and ensure a smooth flow of data.

TWCC, on the other hand, is a congestion control mechanism that works by monitoring the packet loss rate and the round-trip time (RTT) of the data packets in the network. It uses this information to adjust the sending rate of the data stream to avoid congestion and ensure optimal transmission. Unlike REMB, which focuses on adjusting the bitrate of the data stream, TWCC adjusts the sending rate of the data stream.

In summary, REMB and TWCC are both techniques used in WebRTC congestion control, but they serve different purposes. REMB helps the sender to optimize the bitrate of the data stream, while TWCC adjusts the sending rate of the data stream based on network conditions. Both techniques are important for ensuring smooth and efficient transmission of data in real-time communication applications.

评价

这个是比较专门的技术术语, 一开始我没有加上 WebRTC congestion control , 它回答的牛头不对马嘴, 一旦给它加上限定词,它回答得就靠谱多了.

测试三

我又问了一个实操性比较性的问题

Q. how to limit the bandwidth on linux server?

To limit the bandwidth on a Linux server, you can use a tool called "tc" (traffic control). Here's a general outline of the steps:

Determine the interface you want to limit bandwidth on. You can use the "ip addr" command to list the available interfaces and their IP addresses.

Install the "tc" package if it's not already installed. You can use the package manager for your Linux distribution to install it.

Create a new traffic control class using the "tc" command. For example, to limit the bandwidth on the eth0 interface to 1Mbps, you can run the following command:

tc qdisc add dev eth0 root tbf rate 1mbit burst 10kb latency 70ms peakrate 1.1mbit mtu 1500

This command creates a "token bucket filter" (tbf) with a maximum rate of 1Mbps, a burst size of 10KB, and a latency of 70ms. The "peakrate" parameter specifies the maximum rate that can be reached during bursts, and the "mtu" parameter sets the maximum transmission unit size.

Verify that the traffic control rule has been applied using the "tc" command. For example, you can run the following command to show the current configuration:

tc qdisc show dev eth0

This should display the details of the traffic control class you created in step 3.

Note that these are just general steps, and the specific command syntax and parameters may vary depending on your Linux distribution and the version of the "tc" tool you're using. You may need to consult the documentation or search online for more specific instructions. Additionally, keep in mind that traffic control rules can affect the performance of your server and applications, so it's important to test and monitor the impact of any changes you make.

评价

它的回答不错, 尤其是最后的那段注意事项, 它挺明白自己的回答的深度和广度

测试四

最后, 我又问了一个关于拖延症的问题, 这是我的老毛病, 看看它有什么建议

Q. 如何克服拖延症

克服拖延症是一项需要付出努力和坚持的任务。以下是一些可能有帮助的建议:

  1. 确定目标:制定明确、具体、可量化的目标,并在日程安排中安排时间。为了确保目标可行,最好将其分解成更小的子目标,并逐个完成它们。

  2. 创造一个良好的工作环境:找到一个安静、清洁、无干扰的工作空间,避免干扰和分散注意力的因素,例如社交媒体或电视。

  3. 制定计划:制定一个清晰的计划,包括每项任务的时间、工作量和截止日期。把计划写下来,并把它放在一个可见的地方,以便时刻提醒自己。

  4. 立即开始:一旦有任务需要完成,尽可能快地开始,并保持专注和连续性。如果你发现自己开始拖延,尝试告诉自己只需要做一点点,然后逐渐增加动力和专注度。

  5. 分配时间:分配固定的时间段来完成任务,然后设定一个提醒来确保按时完成。例如,将每个任务分配给一个特定的时间段,然后用定时器提醒你何时结束。

  6. 制定奖励和惩罚:给自己一些奖励,以激励自己在规定时间内完成任务。相反,也可以制定一些惩罚措施,例如在规定时间内未完成任务将会失去某种奖励。

  7. 寻求支持:寻求家人、朋友或同事的支持和鼓励。他们可以提供帮助、鼓励和监督,以确保你完成任务。

最后,要记住,克服拖延症需要时间和耐心,成功的关键是坚持努力和实践。

评价

这个回答比知乎的多数回答靠谱, 有了它, 多数时间这种问题我可以不必再花时间再查百度和知乎了

总结

ChatGPT 果然非同小可, 我还做了一个文章修改和扩充的测试, 效果相当不错.

这是人工智能的一个非同凡响的里程碑, 未来它会改变世界的, 如果给它加上耳朵和嘴巴, 长出手脚, 人类会迎来又一次生产力的大解放.


本作品采用知识共享署名-非商业性使用-禁止演绎 4.0 国际许可协议进行许可。