Kids with learning disabilities need AI.
e.g. - a kid has output disorder, it's hard get what’s in their head into their writing. They’re smart, but have trouble expressing it.
AI can help them get the proper teacher feedback to actually grow as a learner.
Here's how 🧵
This is a super cool service! But, I found each AI hallucinated different scientific studies, which made me laugh. But also :(
Great to be able to contrast the results!
We are excited to launch Poe for desktop today at https://t.co/cMuZ6Cb3cY. With this launch, between our existing iOS app this new web interface, we believe Poe is now the fastest and easiest way to talk to AI across iOS and all desktop platforms.
If you don’t feel like paying $20/mo for ChatGPT Pro, try out Poe (by Quora, @adamdangelo). It is currently the only frontend that supports Claude @AnthropicAI, and the ChatGPT interface runs silk smooth. Free for now (at least?)
I like that Poe automatically highlights key terms in the response for you to dig deeper. Click on it and Poe continues with "Tell me more about <keyterm>" on your behalf. Seems like a great way to go down the rabbit hole and boost engagement 100x on Quora 😅
When the price of learning anything is just a few cents, (and way less time per unit of embodied learning) what will you install on your internal operating system?
This temptation - to outsource the 'deciding what to do' part of AI - is going to be dangerous.
Those who know how to interact with the intelligences will be in the best position to protect us - especially those who know when to hold back, and refuse automation.
ChatGPT is reactive: user says X, ChatGPT responds with Y. Risks exist but are bounded. Soon it will be tempting to have proactive systems - an assistant that will answer emails for you, take actions on your behalf, etc. Risks will then be much higher.
If people look to AI to automate tasks with minimum input, there will be a serious competitive advantage for those who are awesome at what we now call 'prompt engineering'
I think then we'll call it 'talking to the AI' instead of 'prompt engineering'
Let's get good at it now!
In LLMs, the need for "prompt engineering" is a sign of *lack* of robust language understanding.
It will be interesting to see if scaling LLMs alone will reduce the need for prompt engineering.
@MelMitchell1 I think this might miss the fact that a better LLM will create new affordances for prompt engineering.
A 'well-prompted' LLM might preform better than an 'autonomous' LLM for all levels of development.
This is amazing.
A professor at Wharton University *expects* his students to use AI and ChatGPT in his course.
The professor states "AI is an emerging skill" and provides tutorials for his students on how to use AI tools.
Here's the course syllabus from day 1:
This is especially erroneous in the area of education - which is largely about assessing process to output relationship.
The process of creation is essential to an appropriate evaluation of the outcome. And an appropriate evaluation is essential for actually growing and learning
Writing ChatGPT prompts feels like managing an intern.
You’ve gotta be pretty specific or you’re going to get something generic and not useable.
Communication as a skill is at a premium right now because of it. Instead of managing humans, you’re managing 175B params on a GPU.
Valid search - across multiple streams of information - is the base of any 'killer' educational ai app.
Right now ChatGPT is a non starter (outside of bing)
I love LLMs but to be a viable search engine, you gotta offer more than just text.
We've incorporated a lot of our apps into https://t.co/UcBQ2GDIxp now in order to be more factual and display data in the best way.
Next week we will improve the design of our citations & links