Comments

You must log in or register to comment.

cartoonzi OP t1_j4puelx wrote

Last week, The Information reported that Microsoft and OpenAI are working on a ChatGPT-powered version of Bing. A use case that raised eyebrows was ChatGPT's ability to answer questions in concise and straightforward language, which seemed like a much better experience than the one we’re used to on Google.

ChatGPT was alarming enough for Google to declare a "code red" and prioritize the release of its own AI products. Google has been building similar large language models (LLMs) but has been much more secretive and cautious about them. You may recall last summer's controversy when a Google engineer claimed that LaMDA, Google's chatbot, was sentient. And Google even built one of the core technologies powering ChatGPT. We don’t know how powerful Google’s chatbot and other AI products they’ve created are, but we know they have the talent and funds to compete.

Currently, Google commands 85% of the global search market, while Bing only accounts for 9%. So Microsoft doesn’t have as much to lose as Google, which explains its higher risk tolerance in releasing a first-of-its-kind product that could be less reliable and provide false information.

But the bigger news is that Microsoft intends to add OpenAI's chatbot technology to its Office apps. Microsoft Office accounts for 23% of the company's revenue, compared to only 6% from Bing ads.

Having generative AI capabilities built into these Office apps will become a competitive advantage for companies that use them. And every company will end up paying for the premium tier of AI-powered Office apps so they can keep up with their competition.

----------

The integration of generative AI into Office apps will completely change the way we work. I can't wait to see how all of this unfolds...

11

lysergic101 t1_j4pueqf wrote

Unless Google go back to the good old days when they actually served up good search results instead of adverts then I believe the impact on Google could be massive.

132

Surur t1_j4pus6t wrote

I kind if disagree. I am already using chatGPT to get answers to questions instead of typing into Google.

28

roughback t1_j4pvsbg wrote

Yup. This weekend I relied heavily on Google to find things I needed, contact numbers for businesses, and get me to places I wanted to go. Not once on my phone did I think "hey chatgpt probably knows where they have this thing I need right now"

When I wanted to know the weather this morning I said hey Google.

7

Paledude4u t1_j4pvv5g wrote

Google sold out to advertisements. It has become stunningly worthless.

2

AadamAtomic t1_j4pw93m wrote

>Unless Google go back to the good old days when they actually served up good search results

They are already working their own EXTREAMLY powerful A.I for that reason.

ImaGEN asking Google to show you a 3D model of what tiger sharks look like, accompanied by correct photos and information without links.

17

FuturologyBot t1_j4px1ey wrote

The following submission statement was provided by /u/cartoonzi:


Last week, The Information reported that Microsoft and OpenAI are working on a ChatGPT-powered version of Bing. A use case that raised eyebrows was ChatGPT's ability to answer questions in concise and straightforward language, which seemed like a much better experience than the one we’re used to on Google.

ChatGPT was alarming enough for Google to declare a "code red" and prioritize the release of its own AI products. Google has been building similar large language models (LLMs) but has been much more secretive and cautious about them. You may recall last summer's controversy when a Google engineer claimed that LaMDA, Google's chatbot, was sentient. And Google even built one of the core technologies powering ChatGPT. We don’t know how powerful Google’s chatbot and other AI products they’ve created are, but we know they have the talent and funds to compete.

Currently, Google commands 85% of the global search market, while Bing only accounts for 9%. So Microsoft doesn’t have as much to lose as Google, which explains its higher risk tolerance in releasing a first-of-its-kind product that could be less reliable and provide false information.

But the bigger news is that Microsoft intends to add OpenAI's chatbot technology to its Office apps. Microsoft Office accounts for 23% of the company's revenue, compared to only 6% from Bing ads.

Having generative AI capabilities built into these Office apps will become a competitive advantage for companies that use them. And every company will end up paying for the premium tier of AI-powered Office apps so they can keep up with their competition.

----------

The integration of generative AI into Office apps will completely change the way we work. I can't wait to see how all of this unfolds...


Please reply to OP's comment here: https://old.reddit.com/r/Futurology/comments/10eb8wq/chatgpt_wont_kill_google_it_will_help_it/j4puelx/

1

angrathias t1_j4py4gx wrote

You need to consider what the question is in the first place and whether it’s provable easy enough. Like If you ask it a programming question, you simply run it, not really different from say stack overflow

1

aurialLoop t1_j4pymnm wrote

Yes ChatGPT will confidently tell you falsehoods as if they were true. There is no fact checking functionality in ChatGPT. It's still incredibly useful and gets a lot right. I've been using it to speed up tasks and do repetitive boring jobs that I would prefer to delegate if at all possible.

22

Tenter5 t1_j4pys75 wrote

GPT will start to incorporate subliminal ads and be shit too.

5

Cyber-Cafe t1_j4q0280 wrote

And googles office products are third rate low tier imitations of Microsoft products. Google cannot even do identity provisioning properly, making their suite useless in a corporate setting. It’ll help Microsoft. It’ll hopefully kill Google.

1

speedshifter2015 t1_j4q07ym wrote

I've noticed chatgpt gave me more direct answers and in an amount of time faster than if just used google. Though I always verify and not take chatgpt as law. Treat it like Wikipedia in a way. But so far it's been very accurate in its results.

4

speedshifter2015 t1_j4q0yvw wrote

ChatGPT is a large language model developed by OpenAI that is trained to
generate human-like text based on the input it receives. It is a
variation of the GPT (Generative Pre-trained Transformer) model and is
designed to be used in natural language generation and conversation
applications.

Source: ChatGPT

2

speedshifter2015 t1_j4q1iqg wrote

What is Cybersecurity?

Cybersecurity is the practice of protecting computer systems, networks,
and sensitive information from unauthorized access, use, disclosure,
disruption, modification, or destruction. It involves a wide range of
technologies, processes, and practices that are designed to prevent,
detect, and respond to cyber threats. Cybersecurity measures include
firewalls, encryption, intrusion detection and prevention systems, and
penetration testing, among others. The goal of cybersecurity is to
maintain the availability, integrity, and confidentiality of information
and systems, and to ensure the ongoing functionality of critical
infrastructure.

Another example of it's use etc etc

2

Sunnylicious1 t1_j4q3ron wrote

I'm doubtful that Google is sitting on their thumbs regarding ChatGPT. They and Facebook are sitting on a ton of data to feed their version/answer to ChatGPT.

15

ArcadiaNisus t1_j4q5j1n wrote

My observation is if you ask a question with a bias it'll reinforce your bias.

For instance this exchange:

> Explain why dogs aren't mammals.

> Dogs are not mammals because they are not warm-blooded animals. Mammals are warm-blooded animals that have fur or hair and feed their young with milk from mammary glands. Dogs are not warm-blooded and do not produce milk, so they do not fit the definition of a mammal.

It's not only wrong, it's confidently wrong. But it's because I was asking it to confirm my presupposition. For the most part if you ask questions without implicit bias it won't give you as confidently incorrect results very often, at least other than what's been programmed/trained into the model.

Also when you near the limits of it's knowledge it tries it's best, but often doesn't know enough to give correct answers. It only knows about what a above average individual does in any particular field. Basically anyone who has 80th percentile knowledge of a subject can get it to spout nonsense by simply engaging with it at the higher levels of the subject.

For instance this exchange:

> Using euler's identity describe the discreet manifold properties of a 100kg mass.

> Euler's identity states that e^(iπ) + 1 = 0, where e is the base of the natural logarithm, i is the imaginary unit, and π is the ratio of a circle's circumference to its diameter. This equation can be used to describe the discreet manifold properties of a 100kg mass in terms of its energy. Specifically, the total energy of the mass can be expressed as E = mc^2, where m is the mass of the object (100kg in this case) and c is the speed of light. Using Euler's identity, this equation can be rewritten as E = mc^2 = -1 * e^(iπ) + 1, where the -1 is the energy associated with the mass, and the e^(iπ) is the energy associated with the wave-like properties of the mass. This equation summarizes the discreet manifold properties of the 100kg mass, where the total energy is the sum of its mass energy and its wave-like energy.

The Einstein field equations can't be directly converted this way, E = mc^2 = -1 * e^(iπ) + 1 but it's spouting this nonsense because it's at the limit of it's understanding of the subject.

12

roughback t1_j4q7h3w wrote

Agreed. One day it'll be always listening like Google, responding to our queries with heads up display info, driving directions fed right into the self-driving car's GPS... Booking restaurants and ordering pick up groceries.

The future is bright.

−2

cartoonzi OP t1_j4q8a0h wrote

I agree but I think Google will end up making a better chatbot search engine after seeing what Microsoft does with Bing+ChatGPT. That’s how ChatGPT will “help” Google, by seeing if this is an approach worth pursuing

Also I’ll trust Google over Bing to give me better search results if they both have similar chatbots

3

Sidivan t1_j4q8u6r wrote

This is actually kind of scary. That’s another level of trust.

Currently, when you search for something, you’re looking for content related to your search. These are independent sources, for better or worse, that you can validate based on their merits. For instance, if I search for information about tigers, I can likely trust WWF, Britannica, etc… but if I search for that and AI generates that info, I don’t have any insight into the source. I have to trust that the AI returned accurate, reliable information.

If AI generates a tiger 3D model that has something obvious like 2 tails, that would raise a mental alarm only because you know they do not. What if you didn’t know that? What if instead of two tails, it just made the black stripes too thick, leading you to believe they aren’t as orange? Would you notice that?

Interesting times when we rely on AI generated content as fact.

23

PurpsMaSquirt t1_j4qazem wrote

Bit of a short-sighted take. If they don’t understand what users are searching for, they’ll be serving ads ineffectively. This would be a waste of the user data they have collected over time. So yes, Google absolutely cares that they are getting you what you need (and quickly) so they can tailor ads for you better while ensuring you keep using them.

5

cartoonzi OP t1_j4qb0fm wrote

The premise of the article is that ChatGPT combined with Bing will help Google see how people use, misuse, and if they will pay for a chatbot search engine while taking none of the risk. If chatbots are the future of search, Google will release a better version of it. They have the most reliable search engine after all

0

Danno1850 t1_j4qgnik wrote

Oh god, you’re right. No one is talking about how ads will play into AI. We didn’t do it with the internet but I think this time we have the tech to make AI decentralized. I’m sick of ad based businesses. We should make it easy to find good products and not give bad products the opportunity to buy their way to profit.

6

abrandis t1_j4qgprr wrote

Exactly this, they rely too much on those ads to satisfy Wall st. That's why they're treading cautiously because if you give users a system that provides near perfect answers and they don't go around clicking a bunch of links it kinda breaks you're business model.

I see this as a Kodak Digital vs. Kodak film crossroads, do you create a better search at the expense of ad revenue, because someone will.

13

zenwarrior01 t1_j4qksdp wrote

The author clearly has zero comprehension about the ramifications of OpenAI's technology, even if you are silly enough to pigeon hole it to just ChatGPT in particular. It has FAR broader use cases than merely writing scripts, stories, etc. Search is gonna be absolutely massive... far bigger than Word integration.

2

Deadboy00 t1_j4qmrjg wrote

This is an interesting trend but I’m having trouble identifying general search queries that would benefit from services like ChatGPT.

Can someone provide some examples that are not about generating code snippets or people trying to plagiarize creative works?

1

Surur t1_j4qrx29 wrote

A real-life use today when some-one asked me for advice about buying a mechanical watch as a gift, an area I know nothing about:


Me: when buying a mechanical watch, what should one look for

CHatGPT: When buying a mechanical watch, there are several key factors to consider:

Movement: Look for watches with a high-quality mechanical movement, such as those made by Swiss manufacturers like ETA, Patek Philippe, and Rolex.

Build quality: Look for watches that are well-constructed, with attention to detail and finishing.

Brand reputation: Consider the reputation of the brand, and research customer reviews and ratings before making a purchase.

Functionality: Consider the features and functions that are important to you, such as a chronograph or a power reserve indicator.

Design: Mechanical watches come in a wide range of styles and designs, so consider what you like and what will suit your personal style.

Serviceability : Consider if the brand or model you're looking at is easy to service and if the parts are readily available.

Price: Mechanical watches can range in price from a few hundred to several hundred thousand dollars, so determine your budget and stick to it.


Me: which brand of watch is recommended for between £250 and £300

CHatGPT: There are several watch brands that offer high-quality mechanical watches at a price point between £250 and £300. Some popular options include:

Seiko: Seiko is a well-known Japanese brand that offers a wide range of mechanical watches at affordable prices. The Seiko 5 is a popular collection that offers a variety of styles and features at a very competitive price.

Tissot: Tissot is a Swiss brand that offers a range of mechanical watches at a variety of price points. The Tissot Visodate is a popular collection that offers a classic look at a reasonable price.

Hamilton: Hamilton is a Swiss brand that is known for its high-quality mechanical watches at an affordable price point. The Hamilton Khaki Field is a popular collection that offers a classic look and robust construction at a reasonable price.

Oris: Oris is a Swiss brand that offers a range of mechanical watches at a variety of price points. The Oris Divers Sixty-Five is a popular collection that offers a classic look and robust construction at a reasonable price.

Bulova: Bulova is a well-known American brand that offers a wide range of mechanical watches at affordable prices. The Bulova Moon-Watch is a popular collection that offers a variety of styles and features at a very competitive price.

These are just a few examples, and there are many other brands and models to consider within that price range. It's always a good idea to read reviews, check ratings, and compare features before making a purchase.


Me: which one do you recommend for someone who does electrical engineering

CHatGPT: For someone in the field of electrical engineering, I would recommend the Seiko SKX009. This watch is a part of the Seiko's dive watch collection and known for its rugged construction, reliability and precision. It has a mechanical self-winding movement, a water resistance of 200 meters, and a unidirectional bezel, which makes it ideal for someone who works with electrical equipment and may need to measure time underwater or in other demanding environments. Additionally, the Seiko SKX009 is known for its affordable price point and it's easy to service and maintain.

Furthermore, for an electrical engineer, a watch with a chronograph function could be useful as it allows for precise timing measurements. Watches with chronograph function can be found in the same price range as the Seiko SKX009, and with a bit more budget some options like Tissot T-Race, Hamilton Khaki Field Auto Chrono, and Oris Divers Sixty-Five Chronograph are also available.

It's important to keep in mind that watches are also a personal preference, and you might have a different preference in terms of style, size and features. It's best to try on different watches and see which one feels the most comfortable and suits your needs the best.

3

purplePandaThis t1_j4qwkne wrote

This statement n ones like it are dumb.. not everyone uses search engines to answer a question or question format

2

rngeeeesus t1_j4qxdcg wrote

Google Assistant is already doing many of those things. I'm pretty sure Google has arrived there much earlier than OpenAI but they don't want to potentially cannibalize their business.

1

Deadboy00 t1_j4qxt0e wrote

Thanks for posting that.

Did it help with your search? In the time it took me to read the ChatGPT response I found via google that the SKX009 was discontinued and, while available to buy, is quite pricey on the used market. The other information is somewhat accurate (EE’s using chronographs lol) but the brand info would also be on the first page of a google search.

Also…c’mon! Moonwatches in ‘23!? ChatGPT needs some fashion lessons.

1

Surur t1_j4r111q wrote

> Did it help with your search?

Yes, because it felt like talking to an expert. Googling the specific devices gave the same result as you, but at least you know why they were recommended, which is a good starting point.

2

autimaton t1_j4r4ohz wrote

Google stinks. Paid subscriber / advertisement-driven results. Even though there are extra steps I go to ChatGPT for my queries before Google. AI will only become more sophisticated in producing desirable search engine results, especially for finding products and reviews.

2

garygoblins t1_j4r4udk wrote

Eh, Google productivity suite is much more modern. Office has so much legacy bullshit that it's super slow. I'm not saying excel isn't still the defacto, but pretty much every other office tool is inferior. That doesn't touch on the shit show that is teams. Who the fuck thinks it's a good idea to build a communication tool on SharePoint?

2

pressedbread t1_j4r76ht wrote

Who knows. Most of us aren't getting paid for redditing or google searches right now, but the economy in the next few decades is likely as unfathomable as current chunks of the economy tied to TikToc influencers would have been 2 decades ago.

2

sekhmettheeye t1_j4r94ah wrote

Google is already a useless dumpster fire of ads and shit blogs. My main uses for chatgpt so far have been asking the kinds of targeted questions I used to be able to pose to Google, but no longer can because of the nonsense it sends back to me. If AI can do absolutely anything to help us be able to find information we are looking for online, I'm absolutely here for it.

1

thetransportedman t1_j4raa9o wrote

Google will obviously buy them out. The question is for how much

1

Adognamedbingo t1_j4rdru4 wrote

This is the issue I feel like people ignore when it comes to LLMs.

Yeah it’s without a doubt impressive how it generates sentences and content, but if you’re not very familiar with the subject of a question you don’t know if the answer is “correct”.

So if something like ChatGPT should take the place of a search engine it would still need to tell you where it has the info from.

And then how big of a difference is it then from what we currently have.

Another thing that I rarely see discussed is why would anyone create content if they don’t get any visibility/customers from their content and the model just gives an answer without any credit to the actual creator?

12

Adognamedbingo t1_j4rf036 wrote

How do you determine that a LLM like chatGPT serve better results when you dont know if the information that it produces is true or not.

This is a huge issue, because these bots are flawed in the way that they are convincing no matter how bad an answer they give.

Also these models are trained on an incredible amount og data, but are not able to crawl the web yet, so any new info will not be taken into consideration by the AI when giving answers.

That is not great if a pandemic breaks out, but neat if you want to know about history.

I think it’s too soon to say it will replace Google’s search engine since it does a totally different job.

In terms of including it in app and automating different tasks I think you’re right it will be a gamechanger soon.

4

ninjasaid13 t1_j4rg9kn wrote

>ImaGEN asking Google to show you a 3D model of what tiger sharks look like, accompanied by correct photos and information without links.

I wonder how they're going to add Adverts in that? Are you going have to watch a 30 second commercial everytime you generate a picture? Is the generator going to randomly generate an ad when you asked for something else?

1

Brhall001 t1_j4rged6 wrote

What’s the official link to this site? To many copycats now.

1

Adognamedbingo t1_j4rgiy9 wrote

Microsoft has a Big share in OpenAI so not gonna happen.

Google however will/has build their own.

And like with search the question will be who delivers the best search results.

Search is Googles core business where Microsoft is a software company so it’s not likely that ChatGPT will be the gamechanger you think.

It will however automate a lot of tasks which could give Microsoft an edge in software over Googles free alternatives, but search is a whole other departement.

2

Sidivan t1_j4rgjox wrote

That’s a good point about content creation. The way we monetize content today is by consumption. The model measures traffic and assigns value to it. If AI is going to serve up all that content in the form of it’s own content creation, who gets paid? Does everybody it references or considers get a slice? In the above example about tigers, does every single one of the millions of photos it referenced get paid or are they simply inspiration? This battle for revenue is happening right now with visual artists and AI art.

2

awhhh t1_j4rjmdf wrote

It’s not totally Google’s fault. People try and game search rankings by building entire sites towards them. Affiliate marketers and other SEO degens did their darnedest to fuck it up and keep Google switching algorithms

2

Wartz t1_j4rjyti wrote

Microsoft is already doing this in their office 365 product line (if you use the subscription).

It can generate all sorts of content on the fly based on context of your existing text.

1

MightyOm t1_j4rl84b wrote

That's because Chat GPT isn't web enabled right now. What if you asked ChatGPT those same questions and it was web aware? Not only would it plan all that for you, it would do it creativity and probably offer you better alternatives. I don't think Google as it stands is enough to defeat an internet aware AI

1

NateCow t1_j4rn4hq wrote

>My observation is if you ask a question with a bias it'll reinforce your bias.

This. The first thing I told it to do (I work in visual effects) was explain why AI will never replace rotoscope artists, just for the lol's. It confidently made a lot of great points that I would personally argue in support of human artists. I then flipped it around and it confidently declared artists are no longer needed.

I also tried having it make an about page for my website, seeing as I've had a website and public profiles on various websites for a long time. It very confidently went on and on about how I'm a full-stack developer with lots of experience in various technologies.

I am not a full-stack developer. I've dabbled in coding and building some automation tools related to my own work, and a little web design, but that's it. I then told it to highlight my visual effects career, at which point it touted how I'm experienced in Houdini and Maya, two applications I have never touched.

Granted, the pool of information surrounding me was probably limited in terms of lots of corroborating data points. It was amusing nonetheless. I'm enjoying using it to generate little blurbs of website copy for a site I manage on the side.

3

NateCow t1_j4rnble wrote

Honestly I figure it's probably just stealing a bunch of stuff from Stack Overflow anyway. I was trying to figure out a regular expression the other night. All my Stack Overflow findings were really complex and I couldn't get any of them to work. Tried ChatGDP and it also spit out a super complicated one that.. didn't work.

1

NateCow t1_j4rnsfd wrote

The main use I want AI for is so I never have to open my email inbox again, like "Her." Just sort through the junk, find the important stuff... just manage my digital life with email, calendar, etc.

0

ddaaddyyppaannttzz t1_j4rrax7 wrote

I agree 100% with this major issue. I did a ChatAI search on topics I know lots about (research and expertise) and it brought back some useful and correct information but also some false info from probably dubious sources We can’t put too much trust in this tech with out knowing the source of the info At least with a search engine you can skip the known suspect sources or at least read and compare the sources. Right now people are putting too much trust in ChatAI and similar, that should change with time (i hope)

2

outtyn1nja t1_j4rwnui wrote

If it kills profits for Mega-conglomo-corps then the PEOPLE CANNOT HAVE IT.

1

SaberHaven t1_j4s5zpe wrote

As someone who works in AI, this is like saying "motor engines won't kill carriage services. Their biggest impact will be farms". In reality, motor engines obviated horses entirely, changed the face of everything from sport to how we wage war, and every industry in between.

Modelling language reliably means we can intuitively define problems for computers, and literally any functional aspect of life can be integrated with this.

2

SecondAlibi t1_j4skqwb wrote

I think a small caveat against ChatGPT in this sense is that its answer is biased by its training and it only offers a single concise response to a question, whereas Google returns a list of results which the reader can pick and choose which one appeals to their own biases.

3

BlurredSight t1_j4smdar wrote

People forget Google Lamda is a thing, they've kept it restricted, and the few people who have gotten to access it have only said the most positive things especially when it comes to creative thought. Zoe Kleinman from the BBC said she got 30 minutes with it but this was a month before ChatGPT was publicly released and I haven't been able to find if she was able to compare her experiences between the two.

6

Mash_man710 t1_j4sqq78 wrote

If this is what is out in the public domain, imagine what the real cutting edge stuff is like behind closed doors..

2

tomistruth t1_j4srlxf wrote

I doubt it. The number of google searches have dropped significantly ever since I started using chatgpt for research and learning. It is just so much quicker to get contextually strutured knowledge about a problem, instead of shifting through tens of low effort blog posts and forums looking for an answer.

If there is a tool that allows them to acquire knowledge 10x faster and that can adapt, simplify and expand on topics according to the user, than old search will simply soon be left behind. No self assurance blogpost paid by google's PR department will change that fact.

11

Chogo82 t1_j4stf3w wrote

Once ChatGPT is released into the wild, I’m curious how long it will take before it becomes a hate spewing Hitler loving Nazi bot.

1

Specific_Main3824 t1_j4tdgbi wrote

Google either needs to buy out ChatGPT or create their own and incorporate it into their search algorithm. They need to move extremely fast, weeks, or months at worst. Otherwise, Google can join Yahoo on the scrapheap of tech that once mattered. I can imagine ChatGPT has been getting offers left right and centre to buy it out.

4

dlepi24 t1_j4tqgu5 wrote

Power automate is a good early example I feel. With some creativity and basic understanding of M365 products you can create some amazing business workflows. You can now just describe your ideal workflow and it will help design it fairly well and then you just fill in the gaps. It will even write custom expressions for you if you give it an example of what you want it to do.

1

dlepi24 t1_j4trm9n wrote

Drafting business emails is pretty good, or templates/SOPs. I've had it write entire BCDR plans, MFA enrollment notices, etc.

It's great for quickly getting some code going for easy to medium tasks, or seeing a different approach to a problem can open other possibilities.

I've actually had it tell me more about some api endpoints than their online KBs have documented.

It's best used as a tool to aid you rather than thinking it's going to replace you.

2

Invisible_Pelican t1_j4uhamm wrote

Articles like this are extremely short-sighted. Satya Nadella has gone on record saying they will incorporate OpenAI into ALL of their products, meaning MS will be punching out in every direction not just one or two. MS will absolutely try to kill or at least heavily damage Google, and it's delusional to think otherwise. Google needs to come up with an answer to this fast if they want to stay relevant.

1

B_lintu t1_j4uhpyp wrote

ChatGPT won't help google. It will help bing and if google doesn't make something similar, bing will overtake.

1

94746382926 t1_j4v5pcg wrote

They've had models that are competitive if not better than GPT for about a year now. They can't buy out ChatGPT since Microsoft basically owns it at this point.

It also wouldn't make much sense since they already own deepmind which is arguably a better AI company. They have the capabilities, they've just decided to keep their cards close to their chest for the time being.

2

LivingHighAndWise t1_j4vxag6 wrote

I think this article is very short sighted. Even in it's early state, I am already using ChatGPT for things I would have traditionally used a legacy search engine like Google for. When I need a direct answer for something that doesn't require or rely on a persona opinion, I can just ask ChatGPT and it gives me a direct answer without having to sift through paid results and ads. Why would use a Search engine over ChatGPT - especially when V2 comes out which will trained on 100 times more data and have access to the public internet?

1

aurialLoop t1_j5949cz wrote

People should be cautious of any answer it gives you where you don't already know the correct answer.

Where is it really useful? When you're asking it to perform either of the following:

1 Perform tasks that don't require knowledge of facts. E.g.

"Reword this email so that it is clearer"

"Reduce the length of this speech while still making the main points"

"Use the above strategic company goals and suggest ways my team who performs the following tasks can help realise those goals"

"Take the following ideas and write a poem that conveys their juxtapositions"

​

2 Perform tasks that you could do (and already have the required knowledge to evaluate yourself), but don't have time to do. E.g.

(As a computer programmer who knows how to read code): "write me a c# class that keeps track of objects in an environment"

"Write the scaffold of a business case for an automatic potato mashing machine"

"Suggest some game mechanics for a multiplayer card game where you have to work together to destroy an asteroid on the way to earth"

1