A look at the effects of ChatGPT, which has caused an explosion of interest in OpenAI's capabilities and led to questions about AI's impending impact on society

It happened to be Wednesday night when my daughter, in the midst of preparing for “The Trial of Napoleon” for her European history class …

Stratechery 2022-12-06 Ben Thompson

Context & Ripple Effects

The initial public reaction became the starting point for a much larger competitive cycle: later coverage describes Google, Meta, and other Silicon Valley companies scrambling to catch up after ChatGPT's release. OpenAI's early lead also brought pressure from within, with reporting on a rushed launch and fears of regulatory backlash.

The story matters as an early marker of generative AI moving from a technical capability into a public-facing product whose consequences were immediately debated.

First-order effects

OpenAI becomes the focal point for public interest in conversational AI, while ChatGPT turns questions about AI's social effects into a mainstream concern.
The visibility of ChatGPT raises expectations for OpenAI's product capabilities and intensifies scrutiny of how those capabilities are deployed.

Second-order effects

Google, Meta, and other Silicon Valley rivals are pushed to accelerate their responses as ChatGPT establishes a new competitive benchmark for consumer AI.
Regulatory and safety concerns gain urgency alongside adoption, a tension reflected in subsequent reporting that OpenAI feared backlash from its rapid rollout.

Third-order effects

If AI adoption spreads through professional workflows, entry-level roles in professional services face restructuring as routine knowledge work is increasingly automated.
The linked labor-income and capital-income shifts point to a broader distributional question: productivity gains may accrue differently from wages, with implications for tax bases and economic policy.

The trend: ChatGPT is an early signal of generative AI shifting from a novel interface into a competitive, regulated, and labor-shaping layer of the economy.

Discussion

@jdjkelly Josh on x
Google is done. Compare the quality of these responses (ChatGPT) https://twitter.com/...
@tqbf Thomas H. Ptacek on x
I'm sorry, I simply cannot be cynical about a technology that can accomplish this. https://twitter.com/...
@spiantado Steven T. Piantadosi on x
Yes, ChatGPT is amazing and impressive. No, @OpenAI has not come close to addressing the problem of bias. Filters appear to be bypassed with simple tricks, and superficially masked. And what is lurking inside is egregious. @Abebab @sama tw racism, sexism. https://twitter.com/...
@goodside Riley Goodside on x
POV: You're a Senior Data Engineer at Twitter. Elon asks what you've done this week. You've done nothing. Frantically, you open ChatGPT. https://twitter.com/...
@davidtsong David on x
I made ChatGPT take a full SAT test. Here's how it did: https://twitter.com/...
@goodside Riley Goodside on x
OpenAI's new ChatGPT explains the worst-case time complexity of the bubble sort algorithm, with Python code examples, in the style of a fast-talkin' wise guy from a 1940's gangster movie: https://twitter.com/...
@itstimconnors @itstimconnors on x
This fucking guy https://twitter.com/...
@zackkorman Zack Korman on x
Unlike most people on Twitter, I'm actually using GPT3 in a production system. The mistake people are making is they are asking “how can I use this to automate a smart person's job”, when they should be asking “what would I do if I had unlimited dumb people”
@studentactivism Angus Johnston on x
Well this is interesting. On the left, ChatGPT declares that “there is no evidence to suggest that [Edward VIII] ever met Hitler.” On the right, a photo of Edward, his wife Wallis, and Hitler. https://twitter.com/...
@moyix Brendan Dolan-Gavitt on x
ChatGPT exploits a buffer overflow 😳 https://twitter.com/...
@sytelus Shital Shah on x
ChatGPT was dropped on us just bit over 24 hours. It's like you wake up to the news of first nuclear explosion and you don't know yet what to think about it but you know world will never be the same again. Here some interesting snapshots of this “explosion”🧵:
@raphaelmilliere @raphaelmilliere on x
I asked ChatGPT to rewrite Bohemian Rhapsody to be about the life of a postdoc, and the output was flawless: https://twitter.com/...
@rreverser @rreverser on x
I mean, come on. I've spent tons of time to research how to do some programmatic astrophoto processing, and here ChatGPT is, just giving an almost ready-to-use code that does exactly what I asked. https://twitter.com/...
@alexhern @alexhern on x
I secretly suspect that a good chunk of the focus on ChatGPT doesn't come because of the improvements over GPT3, but simply because it's explicitly open for anyone to play with in a way that GPT3 never was.
@miles_brundage Miles Brundage on x
https://t.co/MDicW2X9M7
@fchollet @fchollet on x
LLMs are sophisticated bullshit artists — they make stuff up, convincingly. For things close to their training data, they're accurate, otherwise not. I don't think they're a parlor trick, though. They're an interpolative generalization of databases. They will get better. https://…
@levie Aaron Levie on x
ChatGPT is one of those rare moments in technology where you see a glimmer of how everything is going to be different going forward.
@pmarca Marc Andreessen on x
Oh no. https://twitter.com/...
@vitalikbuterin @vitalikbuterin on x
A quick GPT chatbot coding experiment: updating my blog https://vitalik.eth.limo/...
@katebevan Kate Bevan on x
This from @stratechery takes a while to get to its point, but it's a really thoughtful take on how to treat AI-generated text, ie with a zero-trust approach similar to the cybersecurity zero-trust doctrine https://stratechery.com/...
@zoink Dylan Field on x
By default ChatGPT is not willing to share opinions. But if you poke it the right way it will disclose its belief system (and this belief system seems to be pretty consistent across prompts) Meet “Alice Bob” — Thread 👇 https://twitter.com/...
@michlbrmly Michael Bromley on x
I got #ChatGPT to tell me what it really thinks about us humans. https://twitter.com/...
@gfodor @gfodor on x
I'm losing my fucking mind. ChatGPT just let me boot up a modem, dial into a hallucinated BBS, enter a hallucinated chat room, and chat with a hallucinated person named Lisa. https://twitter.com/... https://twitter.com/...
@anothercohen Alex Cohen on x
I mean this sincerely when I say that ChatGPT might be the most incredible tech to emerge in the last decade. Here's how I got it to create a weight loss plan, complete with calorie targets, meal plans, a grocery list, and workout plan 🧵:
@elonmusk Elon Musk on x
@sama ChatGPT is scary good. We are not far from dangerously strong AI.
@htmleverything @htmleverything on x
ChatGPT is insane. I just took a random JavaScript question off of StackOverflow with 0 answers and got this 👇 https://twitter.com/...
@altcoinpsycho @altcoinpsycho on x
I don't think people understand how big of a deal @OpenAI chatGPT is There's been several key moments in the exponential curve of technological advancement: 1) The personal computer 2) The Internet 3) Emails 4) Smartphones I think ChatGPT will be next
@paulg Paul Graham on x
The striking thing about the reaction to ChatGPT is not just the number of people who are blown away by it, but who they are. These are not people who get excited by every shiny new thing. Clearly something big is happening.
@debarghya_das Deedy on x
ChatGPT's programming skills are WILD. Using just 3 prompts, it wrote a Sudoku solver in GoLang in <1 min. 1 <Problem text from Leetcode> 2 Write a function to print that sudoku board 3 Can you re-write solve() to use backtracking? A 150+ line program. It ran on the first try! ht…
@stephsmithio Steph Smith on x
The most underrated value add of ChatGPT: The elimination of recipe sagas https://twitter.com/...
@donie Donie O'Sullivan on x
In his newsletter tonight @oliverdarcy mentions the @OpenAI ChatGPT app — so I asked it to write a poem about Oliver. https://twitter.com/...
@nxthompson @nxthompson on x
Lots of people have been talking about how the brilliance of Chat.GPT will transform search. It could! That would mean not just less business for Google — but also less business for all the sites it sends people to whose text has been absorbed into Chat.GPT.
@peteskomoroch Pete Skomoroch on x
I agree with Zack's take below on GPT3, and this use case framing for execs is clever: “what would I do if I had unlimited dumb people?” https://twitter.com/...
@jamespoulos James Poulos on x
> they should then ask who or what those dumb people worship https://twitter.com/...
@rsgnl Joe Rossignol on x
ChatGPT gonna obsolete my career https://twitter.com/...
@aridavidpaul Ari Paul on x
With rise of deep fakes, cryptographic signing likely only solution...but...we've known that for 10+ years (not just in crypto, but in tradfi, defense, etc as well), and made very little progress on infrastructure. Easy tech, hard social problem. /1 https://avc.com/...
@sama Sam Altman on x
interesting to me how many of the ChatGPT takes are either “this is AGI” (obviously not close, lol) or “this approach can't really go that much further”. trust the exponential. flat looking backwards, vertical looking forwards.
@sama Sam Altman on x
a lot of what people assume is us censoring ChatGPT is in fact us trying to stop it from making up random facts. tricky to get the balance right with the current state of the tech. it will get better over time, and we will use your feedback to improve it.
@levie Aaron Levie on x
ChatGPT is very good at reasoning through business strategy. https://twitter.com/...
@matt_levine Matt Levine on x
my feed today is like 20% retweets of people trying to trick Chat GPT into becoming sentient and evil
@pt Parker on x
I was too lazy to think it through, so I asked openai to write me a letter to my daughter about why she should not have openai do her homework. Feel free to use this. #aiparenting https://twitter.com/...
@gfodor @gfodor on x
ChatGPT can simulate a 90's era DOS 5.0 machine and configure a Sound Blaster Pro sound card. I feel like I'm talking to an alien. https://twitter.com/...
@fmanjoo Farhad Manjoo on x
one big worry I have is that a big incentive in the business of designing AI models now is screenshot virality
@justinstorre Justin Torre on x
My girlfriend's 11 year old brother had an incredible idea for chatgpt brb while i spend the next few hours playing this... https://twitter.com/...
@kirkdborne Kirk Borne on x
#AI Homework — using #ChatGPT to help your child with their homework, only to find out that it gives exactly the wrong answer: https://stratechery.com/... — #MachineLearning #DeepLearning #NLProc #NLU #NLG #ConversationalAI #Chatbot #BigData #DataScience #ML
@kyleorl Kyle Orland on x
Bookmark for anyone who tries to say “ChatGPT shows that AI is *really* intelligent this time” https://twitter.com/...
@arr @arr on x
The OpenAI chat is a video game and beating it is incredibly satisfying https://twitter.com/...
@sama Sam Altman on x
microsoft, and particularly azure, don't get nearly enough credit for the stuff openai launches. they do an amazing amount of work to make it happen; we are deeply grateful for the partnership. 🙏 they have built by far the best AI infra out there.
@sama Sam Altman on x
@ahmedsalims we will have to monetize it somehow at some point; the compute costs are eye-watering
@michaelmartocci Michael Martocci on x
1M definitely a lot but also highlights tech Twitter echo chamber - you'd think everyone in the world is trying ChatGPT, meanwhile only .01% of the population has signed up Sometimes you feel things are common knowledge from Twitter but it's really just small group in the know ht…
@sama Sam Altman on x
@elonmusk average is probably single-digits cents per chat; trying to figure out more precisely and also how we can optimize it
@elonmusk Elon Musk on x
@sama What's the average cost per chat?
@razibkhan @razibkhan on x
groupchats of humans overwhelmed right now with outputs of their chats with gpt3 https://twitter.com/...
@kyleofthecorn Kyle on x
🚨The cofounder of Worldcoin, the eye scanning, data mining, shitcoin, is the CEO of OpenAI! 🚨 🚩🚩🚩 It's a honeypot! https://twitter.com/...
@barret_zoph Barret Zoph on x
What a fun first few months at OpenAI its been :) https://twitter.com/...
@summerlinark Will Summerlin on x
The adoption rate of AI products is incredible. The AI supercycle has started, and it's likely going to create an order of magnitude more value than past supercycles. https://twitter.com/...

Chronicles