58% of US adults have heard of ChatGPT; 14% have tried it
pewresearch.org
pewresearch.org
I would say both of these numbers are really impressive, not just the first one.
Unlike some of those, ChatGPT seems like it has real staying power. I'd love to see their retention numbers.
Moreover, chatgpt is a service offered by - while far from a household name - an established player. OpenAI was founded in 2015. They got their 1 million users in 2022, after 7 years, not 5 days.
And lastly, there’s over a decade between chatgpt and the newest player on that list
https://www.similarweb.com/top-websites/japan/
47th in the US:
https://www.similarweb.com/top-websites/united-states/
I kind of didn't think that ChatGPT was getting traction outside of tech circles but this 100% changed my mind.
The lawyer even said they used it at work. I asked them if their firm had any rules around what they could put into it, and they said no.
That sounds like your lawyer friend might get sued soon!
I am on the faculty of a university in Tokyo, and next week I’m going to be giving a couple of online workshops for the staff members in my department. I plan to show them how they can use GPT-4 in their work: translating and summarizing documents, writing and polishing e-mails, composing Excel functions, brainstorming, etc.
Interest seems high, but usage is still limited. I sent out a survey yesterday to people planning to attend the workshops. One of the questions was “How much have you used ChatGPT, Google Bard, or other LLMs?” Of the 26 who have replied so far, only two replied “frequently.” Half have used them “only a little,” and 11 have not used them yet at all.
[1] https://asia.nikkei.com/Business/Technology/OpenAI-CEO-vows-...
[2] https://www.kantei.go.jp/jp/101_kishida/actions/202304/25shi...
[3] https://www.nikkei.com/article/DGXZQOUC1177N0R10C23A4000000/
I think it was 1995 when WOW!, which was a subsidiary of CompuServe, offered unlimited hours for $15/month. AOL followed shortly thereafter. Prior to that it was something like $4-6/hour, usually at only 14.4kbps - 28.8kbps.
Going from 2.5 - 4 hours a month to a household being able to easily surpass 100 hours a month (for households with multiple kids, easily 150-200 hours a month!) was an absolute game changer for internet adoption, at least in the United States. Then add in Mosaic (eventually Netscape Navigator) taking off and that's when adoption really ramped up. Then cable modems and DSL starting to become more widely available in 1997-1998. Plus instant messaging software basically being the precursor to what we would see later on with SMS.
I thought AOL was the first to offer unlimited hours for a fixed fee, with everyone else imitating it. Am I mistaken?
WOW! was fairly poorly advertised. We originally got it via a free CD at a store (Barnes and Noble or Office Max?), and they started doing TV advertisements a few months later. Though almost immediately thereafter AOL started their giant Unlimited Hours advertising blitz. Marketing truly won in that scenario. CompuServe had a shot, but no one knew about WOW! so it died in obscurity and under the weight of AOL's genuinely gargantuan marketing efforts.
Back in 1995, my local libraries did not have internet access available. This was a city of 100K a fair ways (about an hour) outside of Boston.
School? Super rare. Maybe ONCE OR TWICE A YEAR we would have a class scheduled in the Media Lab where we would learn about what the internet was, and then be able to barely poke at it in a highly supervised manner. This was despite the school having a few ISDN lines at the time, but that was mostly for backoffice administration, with only the one media lab having access outside of it, with a whopping TWO computers for a student body of 4500.
So even in a school with internet access, you might only get to get a vague idea of what the internet was ONCE A YEAR.
1998? Every high-level classroom in my high school of 4500 kids had two or three computers in the back with internet access. School added 8 computer labs (with 40 desktops each) that could be scheduled as frequently as weekly by a given teacher. If I recall correctly the school had upgraded to two DS3s from different providers.
I racked up ~$300 in one month of hourly dial up internet on my parents credit card at that age in that era. Huge amounts of trouble.
Mainly using Kali (software for emulating an ipx lan over the internet) and playing warcraft 2. Most games didn't have internet play as an option so you had to resort to things like Kali.
Irc was fun too but all in all it was nothing like you know it today and it was a pain to get on and use.
Very possible this stat has increased in the last two months, would be interested to see the trend here
Another lesson I might draw is that I need more burner phones, but it gets tedious to queue up for 20 minutes to activate your prepaid SIM via a government ID in this democratic country. Had to verify six numbers for work once because a BigCorp which shall not be named couldn't internally arrange accounts for us to do the desired audit on their systems. Annoying, but we billed them per hour at security consultancy rates (and the client apparently found that worth it) so that's not the same as having to do this to use an online bot.
On the other hand, these third parties might also muddle any survey results that ask about ChatGPT specifically, when people use something like a virtual career coaching site without realizing this was "Open"AI they were using.
I did need to put some effort into configuring TOTP 2FA for the Google account without first giving Google a phone number, but I was able to do it by simulating a WebAuthn device with DevTools, which satisfied the pre-requisite of an existing hardware token or phone number method of 2FA that I needed to gain the ability to enable TOTP 2FA.
You also can't keep signing up for a Google account with the same phone number. It'll say it has been used too often for a few days iirc and then back off again (not sure how often you can repeat that). Otherwise that could be a usable proxy for hiding the phone number from the target site, depending on whether Google gives that information to the third party you're signing up for.
Some of my old Google accounts also don't have a phone number attached, but using that for signups in various places would be an identity as stable as having a phone number.
More generally I'm just getting tired of corps requiring phone numbers. If only there is something like SimpleLogin but for phone numbers (generating single purpose numbers for each source) to prevent phone number collection and spam.
That still is huge though. 500k paying American users alone.
Fwiw, I think people's experiences may vary to be willing to pay. I find the slowness of GPT-4 to mostly offset the intelligence gains on net - hasn't been a game changer for most applications (and bing covers the rest)
The CTO where I work sent a report recently telling everyone in the company the engineering group is now 95% more productive since we've adopted ChatGPT-4. It's absolutely false. I'd say it was actually a negative distraction for a long time and now, we're finding people are using it to abuse our service. So that's a whole new problem we have to try combat.
I have no idea why this tech makes people want to make statements like this but it's just lies. Most people in the team have stopped using it, use it for silly things and jokes /memes, and some just never touched it at all.
There was a thread here with several pro ChatGPT-4 users complaining it's deteriorated a lot, so which is it?
It's nice people find it helpful, but the blanket statements people make about productivity boosts just seem like nonsense.
I don’t know what to say about your CTO. People are definitely overestimating how useful GitHub Copilot is at least (if they say 2x they’re wrong). But there is no doubt that these products are changing the world and making us more productive.
The only use I've found for that so far is giving it a specific URL to use so I don't have to copy/paste some piece of documentation/etc.
I also never type the first prompt out to 4 in the actual chat window if I think there is going to be some back and forth. I write it in a separate file and really clarify my own thoughts before pasting it in and starting the dialog.
If using 4 correctly I think your head should be swimming with new information before you hit the limit for most tasks.
There might be some value in sentiment analysis. But that would mean access to user input and the GPT output data. Thankfully probably available in USA. Just get what people are asking or what they are being recommended evaluate how well they follow it and then exploit them for that money, either to front run or do opposite.
https://www.cnbc.com/2023/05/25/jpmorgan-develops-ai-investm...
School requires multiple paragraphs per student.
In some situations, there are social norms to be respected, but often the content rather than the form matters. So not a condolences card from your employer's HR department, but a report on how a student is doing and how to potentially improve. Does it actually help to write five keywords and have GPT fill it out to such a crazy long report card with no meaning beyond five keywords?
Edit: Or do the teachers just rant a bunch to GPT (or feed it other kind of data, like from a report log like my old high school had) and then ask it to nicely format, professionalize, and remove redundant points? There I would see a beneficial use.
GPT, even 3.5, is very, very good at putting new wrappers on things. It's much better at that than at solving logical challenges or acting as an oracle. If you need that for work, you'll probably already be among the adoptees.