Can Google AI Improve your Workflow?

[@ogjames21]

HS English teacher here: around 14:00, I hear two sides of a major argument around GPT right now. Timely, as I’m grading papers now, too. I think both sides are valid, but I REALLY appreciate the response to Marques’ point. That’s the trick; it IS a tool, but if the tool takes over the work necessary to actually gain mastery (as relevant to the specific task and goal), that’s where there’s an issue. Is ChatGPT currently a problem in school? Yes. Is it also a wildly amazing possibility if implemented well and if education gets its head out of its ass and recognizes how much good it can do? YES. Maybe we could stop acting like it doesn’t exist and stop fixating on the potential for cheating and start exploring how it CAN be a phenomenal tool, discussing and teaching ethics of use, and pioneering innovative ways for educators AND students to utilize it? Not sure this comment makes total sense; my eyes are swimming from the grading

[@Braedon323]

I could see it being useful in teaching students to be skeptical of what they read and fact check. There is a lot of misinformation out there and perhaps with tools like these there will be even more. Even if it generates something that sounds convincing that doesn't mean it is right. Perhaps they could take that as a general lesson in all parts of life.

[@rookiepaw7305]

One AI tool I've been enjoying is Novel AI. On one side of it, you can make it right a whole story given just a prompt. On the other side, which is my favorite, you can take turns writing a story from a blank canvas, where you yourself write a sentence, two sentences, a paragraph, or whatever else, and Novel AI will build on top of it. It will not always write in the way you want it to, and that is okay, because you can change what it says and keep nudging it in the direction it needs to go. As a user of the tool, you must understand the topic, the quality expected, the relevant information, and how to use the tools at your disposal. It's always fun to work alongside Novel AI because sometimes it spits out the silliest things, but also because it speeds up the writing process.

[@ogjames21]

@@Braedon323 I’ve been discussing that with my elective; we’re in a unit on ethics and AI, and we had a pretty cool back and forth on just that topic, as well as a few others. It’s a really good moment to examine that notion of misinformation, especially as companies scramble to pull it into line. We’re getting to see just how easy it can be to blow up truth

[@ogjames21]

@@rookiepaw7305 that’s actually a really cool idea. I’d never heard of it. I’ll try that with them this quarter-thanks!

[@chrislloyd3918]

My biggest concern is for the ESL students I have the writing it spits out is so far above their understanding that they really aren't learning the content or language from using this tool right now. But I agree I'm mostly focused on how to convince them to leave it behind for their benefit rather than punish it's use. It sucks with sources at the moment though.

[@jakeave]

I think using AI for learning will be a game changer because you can ask it personalized questions and it will know what context and what things you do know and don't know and give you better tailored instructions. I think for people that use it right, it will open the world to them in ways that used to not be possible.

[@sarthak-ti]

It's already sort of started on phones. Like on phones, it can learn the apps you open that are recommended to you. But as it gets powerful, it could become more convenient, but also more invasive

[@mattmoore88]

4:29 the conversation about people doing less work. I understand and agree with what you are saying. When we have to work less we learn less and develop the capacity to push ourselves in out of our comfort zone. It will affect mental health over time. If I don’t have to develop the skills of thinking critically we will struggle to express ourselves more and more.

[@AndrewPL5]

If you don't have basic critical thinking skills I don't think you'll be able to use AI properly. For example, I use AI more and more now to compose thing, or basically help me compose things, but you need to understand what to ask and adjust the AI output to what you need as it's not perfect. My boss, however, lacks some critical thinking and he would absolutely not know how to use AI to compose an email or where to start or what to ask. There is many pros and cons to AI, obviously, but I don't think it's as straight forward to what you are talking about regarding critical thinking and such.

[@ipikluninja91]

I actually agree with David about AI leading to people offloading cognitive effort to an artificial intelligence system. It is going to be then tricky when it comes to hiring managers scrutinizing job applications and people are using AI to generate content for them.

[@Mkrabs]

Aren't we doing that already in some way? Offloading "knowledge" into the internet, and only learning how to get informations ? Not much changes, instead of search engines, it"s going to become LLMs

[@esgee3829]

i will just lie and play along with the clarifying question from the hr manager. i'm one with the machine and will go where it takes me. this is for marques' people, in advance.

[@GrantOberhauser]

This is already a problem

[@zulfika_]

And then the generated content has 80% similarity among each other

[@fredrickbambino]

@@zulfika_ not true, it's generated which means it's unique.

[@robertcarolinar]

I agree with what's being said at 4:35. For now, IA it's a tool, a really amazing one but if you don't know how to use it / which prompts are good and if you don't fact check the outcome usually it's gonna be really weird, look like a template and have a lot of errors. For writing in particular, I find it will be really useful for making the skeleton of an essay or a template, saving a lot of time to then use your brain to understand what has been written, fact check, correct grammar errors, etc. I also think it's useful for what Marques said about writing scripts for videos but not using the outcome word by word. It's interesting the possibilities that this tools are about to give us, thanks for uploading this video, it was really interesting.

[@DanielosCompaneros]

I think that many programmers from an era where asm or C programming language were revolutionary, thought that languages with higher level of abstraction or with garbage collection would have the same effect as we think about Machine learning now. I think that programming will become less cumbersome, more accessable and less error prone which is a good thing imo

[@adamoconnor8958]

My thought exactly.

[@FatElk]

Even if it's way worse than Microsoft's AI, it only has to work ok for me to stay with Google Docs.

[@autobahn4610]

Sounds like you're biased toward Google! Just like me!😜🤣😎

[@notyrpapa]

Same - I use MS365 at work but use G Docs on the side whenever I can because MS is so terrible. I WISH Microsoft would put the same effort into UX and product features that they are putting into AI!

[@pxnchx93]

Whats so terrible about Word?

[@JracoMeter]

One thing to consider, if this AI is an assistant it could help summarize or take key points from its own generations. There is definitely more to be done with making it easier for proof reading (one such solution is a highlighting tool for different bits of information in a generation). Adjusting the way you use AI is also important, working in sections instead of long texts or collaborating with it along generation are some potential perspectives. There are so many perspective to consider I often need AI to list a few of them to get a picture of how AI could be implemented.

[@BrazenNL]

I'm with Marques. Getting a search engine to do what you want, is to me, the same as getting an AI to do what you want.

[@cschmitz100]

Teachers listening to you say the student “deserves” a C- instead of an F: 😡

[@username7790]

This was actually a really interesting and information conversation, especially in the second half. Makes me think a lot

[@snelle_tomos]

This conversation is information

[@Lisa_McGuire]

I love how much this podcast cares about our communities interact and intertwine with tech, AI and skill development 💙💙💙💙💙

[@NoOrdinaryScholar]

I'm hoping that with schools it leads to a shift where work is done during classes instead of after school because that's absolutely going to further the disparity between students of different classes. Sidenote: seeing google advertise their ai as a tool to write personalized thank yous to customers and stuff like that felt gross

[@OutOfNamesToChoose]

The combination of copying code from AI and AI written malware sounds like a cyber security nightmare

[@GrellChris]

Co-Pilot will be beat them out

[@rohitkharche7562]

May be AI text generation should also generate Summary of points that the AI has generated with Checkboxs in front of those points as whether this point should be included or skipped. That will solve the proof and accuracy problem.

[@Bmontepeque11]

I'm a computer science engeering student and even though all of the languages already have all of the data structures optimized, and unbelievably easy to use, in my projects the instructions always say "You must implement your own, data structures, don't use the ones from the language" so that we learn how it works! It will probably be the same with AI! Also, the photography example? Your phone does the processing to get great tecnical data, but composition, ruled of thirds, etc. You, the person must know that to make a great photo even if the spec of your photo is perfect!

[@favourphronesis]

Also, I think we need to rethink how essays work in school vs the real world. There are not many professions where you have to know all the information at the top of your head. In most cases, you are allowed, and even encouraged to cross-check your information, and even consult established texts. IMO, That's what AI is doing. It's a much better data researcher, so I can ask it my questions, and then have it spit out all the information it can find, then I organize it till it works

[@ZainTechSavvy]

Loved this clip, lots of grate information!

[@adnansalim2187]

We see a version of chat Gpt writing for humans too. Our fellowship students get professional writers to write theses/dissertations for them with fake data and using softwares to bypass similarity index plagiarism softwares. Those softwares do some real weird crap like renaming gastric acid as gastric corrosive. The people, like in your podcast do not even bother to do a proof read! Furthermore, since the writers aren’t doctors (in our case), the old “database” they tap into to write on a subject (in the introduction, background& discussion sections of their theses/dissertations) are many a times obsolete. For example, while checking a document submitted by some one who had used a professional writer, they referred to a drug (extensively) which had been obsolete and no longer considered a treatment for a disease for many years.

[@crypticTV]

2:50 dude put into words what I have been thinking for years 😂😂

[@tiberiusbaxter7595]

hmm, in return they get access to every single email and thing you write…which they already do but having a.i. suck up everything will progress it EXTREMELY fast in terms of how much information it can articulate

[@JJohnson313]

It's up to the engineers to make it easier to proofread? What are you talking about. The only way to actually proofread is to actually proofread what "you: wrote before sending it. It's the very least that you can do since you didn't actually write it. The other guy was right, that's a user error that you are trying to make an excuse for already. "not my fault I didn't proof read, the machine should have got it right". 04:18

[@guillermocantu8939]

For the first argument about AI writing emails, it’s like Tesla’s self-driving. It’s supposed to help the driver have an easier time driving, but it’s not meant to be the driver. There has to still be someone there ready to take the wheel. So yeah, it’s meant to make your life easier but you still have to guide AI and not let it do all the work unsupervised if you want it to be the way you want it.

[@CarlMaloneR]

I’m surprised nobody mentioned the South Park Chat-GPT episode 😂

[@MireilleLiong]

It's like being a pilot. The lan can fly on it's own but you still need to be in control. So you need to proofread. Always.

[@bokbokeh]

Incredible episode!

[@robertcarolinar]

Really interesting talk, thanks for uploading this!

[@joseaveiga2271]

Having experts in engineering with industry experience would make you podcast more exciting. The conversation would be more guided and intriguing vs just having lots of “I think” conversations . Personal concerns and questions are great but without having any meaningful conversation with an expert, they don’t add much to the dialogue.

[@charlesxavier1904]

I found it funny when in the beginning they said Google is using AI in workspace. Then Marques says "about time" when this has only been used by Microsoft for a week.

[@oru_pavam_nair]

I was thinking why is this clip looking more fresh and happy than other videos from mkbhd/wave form, then I realised that it's the green hoodie that he is wearing. The team should start wearing more colours than red and black and colours in the same palette.

[@Emariess]

Guys, you know you can create an email filter for the word “unsubscribe” and have it skip your inbox and go into a separate “marketing” folder so you don’t have to look at all the spam… I’ve had it setup like this for like 10 years now and it’s literally wiped out all my spam emails.

[@seg5839]

This sounds like having the personal assistant I can't afford. 😊

[@hardikdubal]

I think this is going to put some of the jobs in danger like content writers especially. I also think that shortly there will be applications available in the market first will ask in what tone and way you want to write the content which matches your way of writing email or messages.

[@pwyattk]

Currently internal testing Bard + Gmail integration for drafting emails. Imagine asking Bard to draft an email, but doing it within the inbox.

[@niteshbaskaran2262]

Here a thought. When an AI writes something for you, is it accurately representing who you are as a person in the real world? After all AI will generate similar responses for everybody and very soon most emails will sound the same. Right now, everything written is a reflection of a person's knowledge, character and his language skills. This will no longer be the case if most written text comes from AI

[@chrislloyd3918]

Hold up! Highschool teacher here. The point of having you write an essay in most places is absolutely not getting your answer. The point is the process and the diversity of modes of thinking. Not that every teacher in every educational institution acts on that but that is what it should be about, and that's what's actually beneficial to you.

[@lukereiner]

I'm really curious on a video about privacy, in the AI space/topic of this video as well as your general day-to-day workflow... meaning, do you consider privacy with how surrounded you all are with tech?

[@paulwatrobski8277]

7:26 What about the reverse side where you use AI to summarize a received email? And what if that's an AI generated email? Will the world turn into AI instances talking to each other? An interesting thought. Edit: 20:55 exactly my thoughts 😂

[@ChrisIsOutside]

Joseph Gordon-Levitt is making some good points, but I agree with Anders Holm on this one, I think it's the user's fault if they don't proofread their emails and they will be found out

[@Ostap1974]

We are moving in a terrifying pace towards the future where the number of people who are able to build or fix anything, is miniscule.

[@Dk-qf8dd]

So is the “AI” person specific or a Bernal model that is used by all?

[@kamalrajs9634]

Can i add my windows laptop to devices section in Smart things app or in Google home app?

[@mainkordormawblei]

In a world where we are always busy and need to do so much. We need AI. Now its about asking the right questions.

[@redingtonramos8791]

Plot twist: they'll be using chat gpt plugins 😅

[@TheGreatestJuJu]

Many fear a Super AGI that gains sentience. I’m starting to realize, the really fucking scary thing is mixing Human Sin with a controllable AGI. With the super AGI at least there’s a chance for benevolence and self regulation.

[@michaelasante5002]

I sometimes write code in chat get for interpretation. Its easier learning that way.

[@neoyaku]

This is a good discussion

[@scepter0143]

I don't like where things like ChatGPT is moving, it does your work for you. I think the purpose of AI should be to help humans in doing the thing, like autocorrect and automations.

[@Zodiacman16]

Define "automations", though. Some people would rather everything be automated.

[@scepter0143]

@@Zodiacman16 An algorithm that schedules meetings when the calendars of everyone needed at said meeting are free would be nice. If someone wants everything to be automated for them, then why do they live? The only thing not automated in their life is their heartbeat and breathing. How do they make money if their job can be done by AI? Instead of asking the question "How can AI do this for me," we should be asking the question, "How can AI assist me in doing my normal routines."

[@analog_cafe]

My third blog is five years old, averaging 100 long-form content posts per year. In my opinion, editing takes longer than writing, so I'd rather AI did that better. Writing is the fun part.

[@Peddertech]

Asking AI to write T&C's and us not reading what AI wrote before clicking Accept is how the world ends

[@8stringwrshpleader]

Is there a way for AI generated content to be digitally watermarked so that at the very least it doesn't begin feedback looping false information? I'm sure yall have talked about it before, but if it's writing articles or generating images is there a way to prevent it from pooling from other AI generated information and skewing away from truth or reality as it slowly telephones the information across various generations of AI iteration? Basically, if an AI generated image of a person gets added to its reference pool, could it start adding fingers and warping eyes because it now things that's an acceptable image with respect to its reference pool?

[@Polygonhunter]

So basically I’m writing a prompt that generates a mail, so that the recipient uses an AI to summarize it. Could easily send the promt directly.

[@Salisbury2015]

It will be interesting when it becomes common for corporations to have an AGI as a voting board member.

[@teawa_]

I honestly hate that people are just gonna use AI to cheat, it’s gonna be bad in the long run if these companies decide to start charging for these features. I’ve been using AI to teach me, if I’m unsure about the way im sounding in a sentence, ill send it to chatgpt to correct and tell me where I went wrong, I do the same with code from my classes.

[@Mmartins1097]

Whatever shouldn't the AI talk back to you whenever you ask it to do a task for clarity? Just like when I ask a friend or coworker to help me with a particular task, they ask follow-up questions for clarity.

[@OreoDave]

Nice philosophical debate ❤

[@ismailffahad]

That was a very interesting conversation at the end there. Makes me think about all those "lost techniques" video especially about food. Sometimes the food tastes better the hard way but there comes a point when the economic of it doesn't make sense and so the inferior but easier technique win and over time it gets lost to time. But there is also always somebody who takes that extra effort to "re-discover" that inefficient way to do that thing. Ultimately though through history we will keep losing things but hopefully not to a point where we become the fat people of Wall-E.

[@MegaDenie]

I did exams with chatGPT4 and I got A+ all I needed to do was make it pretend to be a professor responding to another professor. You have to give it a lot of context though for it to give you what you want.

[@themikereda]

Y’all are gonna enjoy ai until it’s reviewing tech and podcasting. Then I bet you won’t be thinking it’s quite as fun. There are definitely some useful ways to use ai but it’s only gonna get better and better, and as an artist and creator, it only looks like it’s going to be replacing creative labor and that kinda sucks.

[@SamzarSalsa]

completely agree

[@cyrilkevin1546]

marques wearing two watches 👀. What could it be

[@TheBoxingNinja]

There's people who still make butter the old fashioned way, just because most will not learn and just use AI, there's still going to be that guy using Windows 7. I don't know how to shave with a single blade, but many do, not most.

[@wenhanzhou5826]

Imagine a word when people dont write their mails and people don't read their mails, but rather summaries of AI's that have been produced by AI's. Yeah, that world is coming...

[@cmanitin13-zi8rb]

22 min Clip WOw

[@marcin4913]

Microphone covering half of the face: choices

[@chad_bell]

today people use computers on a daily basis, but many of us dont know how to write a script to a computer program inorder for the program to function for us to utilize it mindlessly everyday. Like going from solving equations on paper to using a calculator.

[@Riveloperinc]

But if it’s using existing knowledge to build its base, what if people stop adding to that base because they depend on AI.

[@SirMayaba]

Is it me or these Waveform Clips are getting longer?😂

[@Buharialtinee]

That Nice Just like copilot

[@yveslegh]

Let's see where Google AI will bring us!

[@SirWobblySausage]

It's crazy, Ye olde time folks would say that creating things out of thin air was impossible, in fact they'd say it was magic. Magic is only technology we don't understand. Soon we'll be able to speak and things will be done. What we call a prompt, people in the future might say spell.

[@fredrickbambino]

At least it's not going to forcefully reset itself if you ask it about sentience.

[@CyclingGeo]

I tried to ask Chat GPT to write very specific code in SQL for a GEOVIA Surpac macro. Unfortunately it was useless. Looks like I still need to do them myself. Shame I really don't know how to code.

[@marcin4913]

It's fun, unless AI is coming for your job. We tend to enjoy drama when it affects others' lives

[@SUSHI4lyf]

Please also talk about LTT's hacking. 😢

[@meraak1]

who

[@aaron284]

@@meraak1 Linus tech tips, really big Canadian tech chanel

[@knwr]

I mean, people who don't proofread won't make it very far. It's not that complicated or surprising... Those people are genuinely incompetent. Mindlessly copying stack overflow already gets you demoted/fired.

[@GraysonZimmer]

I want to open sheets and ask it to pull my reoccurring subscriptions and put them into a google sheet and give me a sum and dates.

[@therock1232100]

Writecream and Jasper ai are good for what asking

[@jjclaxton1652]

Question: Can A.I. be able to detect false information?

[@OrganicStuff1]

If AI does everything what do humans do what’s our purpose

[@EdwardCulpepper]

Me wanting a smart friend group like this 😂

[@LassiePushedMe]

Who's grading those essays? The teacher also uses Chat-GPT to grade them :)

[@felipin0709]

South Park has an episode about this

[@Wreklessink]

GET THE OTHER GUYS AN EXTRA MIC

[@marcin4913]

The fact that AI can turn a photo of you sitting on the toilet into your wedding photo on a Hawaii beach is frightening, especially when considering that it could make the job of a photographer obsolete in the future. And there's no time to adapt.

[@hastyscorpion]

But the purpose of a wedding photo is to capture an actual moment in time that you know you were a part of. Turning a photo of you on the toilet into a wedding picture doesn’t do that. They aren’t substitutes.

[@levi.s722]

To be honest, i don’t think it will be that useful in texting/answering. Maybe for standard email/Textes about things which are very basic. in my daily business it’s way harder to put all the information necessary in the AI machine, in order for it to create the right answer than to write it by myself. At a certain level the brain has already connected all the data, experience, knowledge, insights and evaluated it. AI has to know all the things happening in the company/team/department to really help you with that. But at that point, when it knows everything, it will be better than all of us. Same with dashboards out of raw data. You learn so much while making a good sheet/dashboard, always find sth you have to deal with, add, delete. I think AI can help you step by step but not with 1 single order. What I really want to have is AI writing protocols for me or creating PP sheets out of existing informations and charts. That save so much time. Or when you have to deal with a new topic/field, to get you a first good overview and how to go on. We‘ll see what is happening.

[@NeverGiveUpYo]

Cool clique ;)

[@juniors3275]

I’m not sure if there’s going to be a time where everyone knows how to code. 🤔

[@mainkordormawblei]

Its a double edge sword

[@VaibhavShewale]

sed, i thought you will try it or something!

[@Brisingr750]

6:21 Someday people will listen to that sentence and laugh

[@Apple-xt4vp]

Her king

[@VictorS72]

❤

[@BMohantyone]

Google have all the resources to win but it can't because of their profitability. Their own business model is killing innovation. But Google is never going down. The whole search engine and answering engine will split between bing and google search.

[@hanstubben]

The film Idiocracy from 2006 is becoming reality! A world full of stupid people depending on a system that does all for them.

[@TheMyrtle77]

Marques' comments on grades does make him come across as someone who has led a very sheltered life, with little insight into education or the necessity of genuine skills outside of tech.

[@lbeetech]

That would be a no...

[@cvasquez601]

I'm sure when calculators first came out people had similar mentality to David. It's always been that you should proof read the work

[@tarzan00745]

Who are those two new younglings there? Why are they there lol?

[@freescape08]

And worse than just the laziness of people avoiding proof-reading, think about how argumentative chatgpt has proven to be...

[@hastyscorpion]

I think you need to check the sources and filtering of your information. Chatgpt has definitely not been “proven “ to be argumentative. Just like go use it.

[@freescape08]

@@hastyscorpion Fair enough, I haven't had time yet. Everything I think I know, I've heard from LMG, and that's only as far as I remember it.