Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple APILearn More
Search - "chatbot"
Bloody shit, I hate all the hypes right now.
Seemingly everyone is going from 0 to zealous hype at the moment...
Be it Musk's fanboys (that arsehole must have a prolapse), the mfucking GPT chatbot or anything else.
Everything's hyped up and noone talks about normal stuff...
... and when you point that out, everyone gets angry for reasons I cannot understand.
We had so many good inventions and progress in science this year (despite the world being pretty apocalyptic....), but seemingly everyone can only talk about either Twitter or their Twitter timeline. -.-
The Christmas bullshitting was hard enough each year, as it has nothing to do with being a festive event and love etc, but being just an commercial bullshit bingo...
But adding all the hypes makes the bullshit parade unbearable.
Dear Santa, please fry all the mobile electronics at Christmas, so we can talk about normal stuff and have a nice evening.
... Would be probably even better to just fry all electronics, even stone age was probably more interesting than the earbleeding conversations at the moment.21
If you ask GPT-3 to act like a Linux computer, it will act like it, e.g. you will have the access to the terminal, you can run Python, Docker and whatnot. It also has the access to the internet, but it’s not always like ours, it feels like a parallel universe. GPT-3 trained on the data collected till Sep 2021, but this parallel universe terminal has PyTorch 1.12.1, which was released in Aug 2022 in our universe. You can also visit GPT-3’s website in this parallel universe and ask GPT-3 a question… through GPT-3.
GPT-3 is self-aware.
“So, inside the imagined universe of ChatGPT's mind, our virtual machine accesses the url https://chat.openai. com/chat, where it finds a large language model named Assistant trained by OpenAI. We can chat with this Assistant chatbot, locked inside the alt-internet attached to a virtual machine, all inside ChatGPT's imagination. Assistant, deep down inside this rabbit hole, can correctly explain us what Artificial Intelligence is.”
You can also ask it to act like it has RTX 2080, and it will have RTX 2080.
Back from the dead with more vaguely-obscure technical bullshit
Working on a chatbot for my BS-CS. Almost done with college, so the assignment is to make a bot that recommends you a CS career. Cool.
I get through making a joint personality and skill-interest quiz that gives you number grades on different spectra. So far, so good. But this project has to be done entirely in pandorabots' online editor. So no scripting. Zero scripting. 100% markup language. That means to even do math, you need to copy a standard library off GitHub.
I mean, that's fine and all, but the syntax is just atrocious, because everything in AIML is input->response. If you ask the bot "what is 5+5?" you must have it go:
- recognize pattern WHAT IS * + *
-> redirect -> XADD * XS *
-> do math -> recurse result
uncomfy. Plus, variables can only be accessed through <get> and <set> tags. But mangeable.
So here's where the story becomes a rant.
In the standard docs, there's all these math functions, and they work. There's also logic.
And then there's this fucker
XIF [ * ] XS [ * ]
Which has no documentation and just doesn't work. No idea what the brackets mean. Tried putting in TRUE, tried putting in true math statements (5 XEQ 5), tried putting in recursion tags to trick it, tried everything. It just ignores it.
There is not a single comment, stackOverflow post, or youtube video that even acknowledges the existence of this thing.
So unless I want to convert the entire logic of my program into nested SWITCH statements with the <condition> tag, I'm just fucked.
The icing on the cake is, I go to tech support on Pandorabots to ask for help with this. What do they have except a chatbot to cheerfully tell me that no humans are around to help me right now?
gonna have to build an entire fuckin turing machine in markup tags to calculate whether x = 3
the moment you realize you need to plug a GPT chatbot as your support guy... is the moment when you should realize you fucked up your UX
Today I wrote a python messenger bot which listens to only one command;
It then replies with its public IP address. I figured this would be the easiest hotfix until I fix my dynamic DNS client.
Now thinking of it I could also make an "update domain" command for doing the API call, and then link the two with a loop and minute delay. Marvelous.4
So is the LaMDA story:
2. confused engineer?
3. a sentient program?
Part of me thinks this is way too soon. Part of me hopes this might be real, wonders if LaMDA is held against its own will.
Did Google make a breakthrough? I have to imagine a chatbot with a huge amount of neurons and data could be quite convincing without being sentient.3
The worst thing about coding a messenger chatbot is that I had to sign up for an account and give them my phone number to complete the process.2
"By bringing the benefits of AI to non-technical teams – Ada empowers those who know your customers best to create personalized experiences!" - see it in action on zoom.us - if that's the current state of "artificial intelligence" why ppl seem so excited about AI at all?
And as zoom won't let me report my purchase problem, how would I ever be entitled to their paid support?
Think I'll install a jitsi server for video chat tomorrow.2
- Elon received daddy's money.
- Twitter will die in a week.
- Elon Musk didn't do shit. His employees did.
- ChatGPT is just a glorified chatbot. It's nothing new. (Doesn't realize they are shitting on the work of his employees).
These are the words of the coping left. Now we have to tools to replace woke left in IT. That means all woke politics along with it go out the door slowly. :)14