What interests me about ChatGPT isn't quite its strengths and weaknesses, but rather the sheer progress that has been made in AI development in the past year. The fact that ChatGPT can write coherent sentences and paragraphs with clear continuity compared to previous AI chatbots is a significant step up. It's development very much has my full interest to see what comes next.
Science Thread
- Thread starter gadgetcool5
- Start date
manqiangrexue
Brigadier
Yeah, don't believe in screenshots with evidence; believe in gadget's anecdotes, like this and these from the past:
"One time, my dad and his friends were in the living room complaining about the CCP. And then suddenly, all their Wechats shut down and Chinese agents came in and beat everybody up! This is crazy, man! China's like crazy Orwellian draconian society, man!"
"Wechat pay got shut down, guys! I can't make any payments; the US government shut it down! I told you guys US is too powerful to mess with! China should plead for better relations!" - Different user: "I just paid 5rmb to my mom. Why are you so bad at lying?"
Last edited:
They made a lot of progress but it's not earth shattering that it will improve my productivity as a software developer.
It might be a good tool for if I forgot to do something in a framework and a chat gpt prompt shows up in my text editor where I can ask for a example of the thing I forgot.
Google developed the Bard chat bot, a direct competitor to the ChatGPT chat bot, which has become incredibly popular. Let's briefly explain what it is. Chatbot Bard and ChatGPT is essentially a regular chat in the messenger. In it, you can ask any questions and receive answers to them in text form. The ChatGPT neural network is a project of the OpenAI team, answering user questions, the ChatGPT neural network is based on its information base. ChatGPT answers can be confused with the answers of a real person, the neural network can write poems, articles, essays, pieces of program code, and so on are already being written with it. The emergence of these networks has been compared to the discovery of electricity, and it has been suggested that these neural networks may soon put some people out of work. According to Google CEO Sundar Pichai, the Bard chatbot is an experimental AI conversational service based on the LaMDA platform. For detailed responses to queries, Google's Bard neural network will use information from the Internet.
Refuse to embrace the latest technology in the name of conserve "tradition"，is the source of backwardness. As can be seen in islamic countries，and countries late to adapt digitalization such as e-commerce and mobile payment. They will have to evolve and adapt it eventually，so it's better early than later. When a new technology emerges，those who embrace it earlier will have an edge
I don't really like this chat bot.
Seems rather biased.
If I say Chairman Mao, the chat bot should respond with some revolutionary spirit and vigor.
Would it even provide a Mao quotation? Those white programmers and Indian programmers will never do that.
Just routine answers to technical and scientific questions.
Don't like it？Come out with a better product
Gender reassignment and its associated artificial hormones / blockers is also a technology. Crypto coins and NFTs are also a technology.
Pursuing technology is all fine and dandy, but let's not pretend every single latest technology is a net positive on society and should be adopted en masse.
I think it's quite reasonable to be conservative with technology whose actual utility is dubious and margin compared with its capacity for harm. With mobile payments and e-commerce, the benefits were quite obvious; lower costs and friction for transactions which would increase the volume and speed of transactions. The concern was security but at least that is a well-defined and understood problem with very little chance of unintended consequences.
With Chat-GPT style "AI"[1]? It's little more than a search engine / chat bot and reveals it's fundamental flaws when asked anything complex. On the coding side, ask it anything more complicated than boilerplate/toy code and it just spits out nonsense. Not only nonsense, but sometimes hard-to-discern nonsense that may fool the less attentive. What is the utility here? Replacement of call-center (live chat) representatives? That is a legitimate reduction of cost for companies, but there are so many more downsides. Producing semi-convincing items that simply don't work. Vast increase in ghost writers answering school questions, essays, etc. Further and more convincing astroturfing on social media and e-commerce websites. The list goes on. There are so many easily foreseeable societal consequences, let alone the unknown unknowns. Societal problems are also, in general, the hardest to resolve. The benefits are not even close to offsetting the costs here.
It is quite easy to predict whether emerging technologies have potential to be useful. It's baked in to its business proposition. That only thing we don't know is how long and how much effort is needed to get that technology to a useful state. For example, fusion is going to be useful if it ever works. The problem there is we don't know how many decades or even centuries is needed to get it to work. On the other side, we also know ANT-style regulation-less digital banking is not going to ever be useful, regardless of how easy or hard it is to achieve. Stuff like Chat GPT and deepfakes in their current form lands in the latter rather than the former.
Research should continue of course, but not mass adoption.
[1]I specifically mention "AI" here because the approach demonstrated by Chat GPT and it's like are fundamentally unable to achieve GAI. Hence, we must be aware of its intrinsic limitations regardless of how developed it becomes. In other words, it can't become more than a chat bot.
Last edited:
Dude, ChatGPT and other large scale language models are coming for your online customer support jobs in 3 years. These AI companies will then also allow external companies to fine tune the model for their specific business/products which will further eliminate customer support jobs
People who aren't into technology don't/can't understand that we have had exponential AI growth in recent years. Its not just about ChatGPT, all other big tech companies have their AI models which they have refused to release because of reputational risks for the product not being ultra-polished. OpenAI's move with ChatGPT has just opened the floodgates.
Baidu, Alibaba will soon follow
What for?
Sorry, I know I will not use it.
What I want to ask, there is no way it will know.
It's backward looking, not forward looking.