OpenAI's ChatGPT Passes One Million Users in Five Days. Is This the Next Big Interface?
SAN FRANCISCO, Dec.
Key takeaways
- •The tool launched on Wednesday as what OpenAI calls a "research preview." You type a question into a simple chat window and it replies in full sentences.
- •It is built on a model from OpenAI's GPT-3 family, further trained with human feedback to make the answers more helpful and less likely to say something harmful.
- •Social media has filled up with examples, some impressive and some strange.
- •Some social media users say a chat style interface could eventually challenge traditional search.
The short answer
SAN FRANCISCO, Dec. 5 - ChatGPT, a free chatbot from the research lab OpenAI, has passed one million users in five days, according to the company's chief executive, Sam Altman. The tool launched on Wednesday as what OpenAI calls a "research preview." You type a question into a simple chat window and it replies in full sentences. It can explain a concept, write an email, draft a poem, summarize a document, or help debug computer code.
SAN FRANCISCO, Dec. 5 - ChatGPT, a free chatbot from the research lab OpenAI, has passed one million users in five days, according to the company's chief executive, Sam Altman.
The tool launched on Wednesday as what OpenAI calls a "research preview." You type a question into a simple chat window and it replies in full sentences. It can explain a concept, write an email, draft a poem, summarize a document, or help debug computer code. It remembers what you said earlier in the same conversation, which makes it feel more like a back and forth.
It is built on a model from OpenAI's GPT-3 family, further trained with human feedback to make the answers more helpful and less likely to say something harmful. Earlier systems from OpenAI and others were available to developers, but few ordinary people tried them. ChatGPT is open to anyone with an email address.
Social media has filled up with examples, some impressive and some strange. Users have asked it to write a Seinfeld scene, explain quantum physics to a child, and write a short essay on the French Revolution. Programmers say it is good at fixing errors in code. Teachers have started asking what it means for homework.
The limits are real. ChatGPT can state wrong answers with total confidence. It sometimes makes up facts, names, or sources. OpenAI says so in plain language on the page, and warns that the tool may produce incorrect or biased answers. It also has limited knowledge of events after 2021. Stack Overflow, a popular site where programmers trade answers, has temporarily banned answers written with the tool because too many were wrong.
Costs are another issue. Altman wrote that the compute costs are "eye-watering" and that the company will eventually need to monetize it. For now it is free.
Some social media users say a chat style interface could eventually challenge traditional search. Others caution that a demo is not a product, and that accuracy problems must be solved first.
Whatever happens next, a lot of people just met their first conversation with a machine that feels like it understands them.
Sources
- 01ChatGPT crosses 1 million users five days after launch, TechStartupstechstartups.com
- 02ChatGPT, Wikipediaen.wikipedia.org
- 03Stack Overflow bans ChatGPT-generated answers, The Registertheregister.com
Corrections
No corrections have been made to this article.
About the reporter
Technology Reporter, Trestlewire
My first published story, in the spring of 2004, was about a database company whose new release had been announced three separate times. I knew it had been announced three times because I had signed off on two of the announcements. I spent seven years as a product manager in enterprise software before I switched sides, and the habit that came with me is a single question: does this ship, or is it a demo.
Read full bio and all stories →The Morning Wire
The day’s reporting, in your inbox before nine.
Free. Unsubscribe any time. We never share your address.