LLM Hacks With Brian Brushwood - DTNS 4922
Update: 2024-12-24
1
1Share
Description
Brian Brushwood explains how he uses AI models in his personal and work life. Plus a Canadian startup has developed a continuous potassium monitoring device that’s inserted just below the skin to help those with chronic kidney disease or at risk of heart failure. And we share our picks for the best holiday gift.
Starring Tom Merritt, Sarah Lane, Brian Brushwood, Roger Chang, Joe.
Link to the Show Notes.
Comments
Top Podcasts
The Best New Comedy Podcast Right Now – June 2024The Best News Podcast Right Now – June 2024The Best New Business Podcast Right Now – June 2024The Best New Sports Podcast Right Now – June 2024The Best New True Crime Podcast Right Now – June 2024The Best New Joe Rogan Experience Podcast Right Now – June 20The Best New Dan Bongino Show Podcast Right Now – June 20The Best New Mark Levin Podcast – June 2024
In Channel
Download from Google Play
Download from App Store
United States00:00
00:00
1.0x
0.5x
0.8x
1.0x
1.25x
1.5x
2.0x
3.0x
Sleep Timer
Off
End of Episode
5 Minutes
10 Minutes
15 Minutes
30 Minutes
45 Minutes
60 Minutes
120 Minutes
Transcript
00:00:00
Daily Tech News Show is made possible by you, the listener.
00:00:03
Thanks to all of you, including Oli Sanjabi, AB Puppy, Dale Mulkehi, and BJRB.
00:00:11
On this episode of DTNS, our tech picks of the year, some tech to help people with harder kidney disease, and Brian Brushwood talks about how he uses LLMs to make his life easier and his creativity better.
00:00:27
This is the Daily Tech News for Tuesday, December 24th, 2024, in Los Angeles eagerly awaiting Santa.
00:00:35
I'm Tom Merritt.
00:00:36
>> From Studio Animal House, and this is not spiked eggnog, I am Sarah Lame.
00:00:42
>> I'm the shows producer Roger Zhang.
00:00:44
>> Hi.
00:00:45
>> Hi.
00:00:46
>> Joining us is the host of World's Greatest Con, Brian Brushwood.
00:00:51
Welcome, Brian.
00:00:52
>> Hey, man, it's good to be here.
00:00:53
It's been a minute.
00:00:54
Bit rusty.
00:00:55
>> Yeah.
00:00:56
>> Hang with me for a bit.
00:00:57
You sneaked in right before the end of the year, so you can check 2024 off the list, then come back and-- >> Actually, that thought actually occurred to me.
00:01:06
>> Excellent to have you.
00:01:07
I can't wait to talk about your usage of LLMs, because I know a lot of people are tired of hearing about the legalities or the controversies, and just want to hear some practical advice.
00:01:17
How you use them?
00:01:18
That's going to be fun.
00:01:19
We shall start, though, with the Quick Kids.
00:01:24
WNASH in our subreddit noted that Reuters sources say Microsoft is looking to rely on its smaller in-house models, like 5/4, rather than open AI models,
00:01:35
to help make Microsoft's co-pilot assistant more efficient.
00:01:39
Smaller in-house models would make co-pilot less expensive to operate, and Reuters also reports that Microsoft is considering partnering with other companies beyond open AI.
00:01:50
Your government of Russia has ordered a stop to cryptocurrency mining in 10 regions of the country, starting January 1st, and running until March 15th, 2031.
00:02:00
So a six-year-long ban, high-energy consumption was listed as the reason for the prohibition.
00:02:05
That follows Kosovo outlawing all crypto mining in 2022, and Angola, criminalizing it in April of this year.
00:02:14
>> Iran's government lifted its prohibitions against allowing connections to WhatsApp and Google Play.
00:02:20
Iranian citizens will no longer need to use a VPN if they want to connect to these services.
00:02:27
>> James, 668 brought us to this sad news on the subreddit.
00:02:31
PC World announced that it's hardware executive editor Gordon Ma-ung died of pancreatic cancer, aged 58.
00:02:37
Ung was famous for his rant of the week segments on the Maximum PC podcast back in the day.
00:02:43
He was also the driving force behind Maximum PC's annual Dream Machine project.
00:02:46
I know a lot of you were fans of that.
00:02:49
Even more recently, he hosted PC World's The Full Nerd podcast.
00:02:53
PC World has a post with a lot of memories from lots of different people that work with them.
00:02:57
We'll have that in the show notes.
00:02:59
>> The information reports that two people with direct knowledge say that open AI has examined the idea of getting back into robotics, and even maybe building its own humanoid robot.
00:03:11
Perhaps related, the company has invested in figure and one X, along with the general purpose AI firm physical intelligence.
00:03:19
Open AI closed its robotic division back in 2021, but maybe not forever.
00:03:26
>> Canada's proton intelligence has developed a product that will provide continuous monitoring of potassium levels with clinical trials to begin this coming year.
00:03:36
Potassium levels are important if you have certain kinds of kidney disease or heart disease.
00:03:41
Proton's device is inserted below the skin, so this is some medical device.
00:03:45
That's something you're just going to pick up off the shelf.
00:03:47
Then it can connect wirelessly to a smartphone, and there's an app that can notify you if your levels fall outside your safe range, and your safe range is determined by multiple things about you,
00:03:58
your weight, your medical history, your diet, your medications, your on, etc.
00:04:04
Also we'll collect information about your potassium levels into a dashboard that you can show to your doctor or other caregivers to help fine tune your therapies.
00:04:12
He was founded by Vancouver's Sahan Rana Mukharachi, who previously founded Microdermics, a skin-based drug delivery startup, and Melbourne's Victor Cardoso,
00:04:23
who is a professor in micro and nano-sensors at Manash University in Melbourne.
00:04:28
The CEO runs the headquarters in Canada.
00:04:31
Cardoso is the CSO, runs research and development in Australia, so it's a multi-national company.
00:04:37
Founders believe their solution is more scalable than others, and there's quite a few others working on tech for this.
00:04:43
Colorado's OLEO monitors potassium levels for dialysis patients.
00:04:48
There's a company called Renalize in Spain, which uses blood samples from a finger prick.
00:04:53
California's a live core, estimates potassium levels based on your heart activity, so it's a little non-invasive, and the UK's Collium Health is recruiting patients for its own clinical study of a rapid potassium self-test.
00:05:07
This would have been something my dad would have been interested, if I know because of his heart disease.
00:05:11
He had to keep track of his potassium levels, but the way he did it was kind of pay attention how many bananas he ate every day, and take potassium supplements and then go in for occasional blood tests.
00:05:22
It wasn't even something he would measure every month, much less every minute of the day.
00:05:29
I think this kind of thing is going to be useful for a lot of people, so another tech-done good story here.
00:05:33
Well, and you know what he reminds me of is, and this is a strange parallel, but in the world of marketing, let's say eventually you want somebody to buy an object.
00:05:42
Every time you require somebody to actively click on a thing, you lose 80% of everyone.
00:05:46
You know, 80% of people won't open your email, won't follow the link, won't look at the product, won't click to buy.
00:05:52
Any kind of active intervention, like the null, the default should be that you get the information about whether or not your levels are correct or not.
00:06:00
The idea of gamifying one's health, nobody would on purpose make sure to swing outside of their zone, but the constant onus of it being on the end user to essentially decide,
00:06:12
"Oh, I need to pay attention to how many bananas I had and estimate all this stuff," as opposed to a one and done medical intervention of having something to monitor it, and you can't help but find out whether or not you're in the green zone or not is going to make it very,
00:06:25
very easy to adjust a lifestyle simply to get the award points for making sure that you stay in the green.
00:06:31
It kind of provides its own dopamine hit, right, just to be in the right range there.
00:06:39
I mean, just for folks who are in the situation where they're doing dialysis, whether it's daily or, you know, regularly, for many folks it is daily,
00:06:50
you know, for something this doesn't sound like, it's not going to keep you out of, you know, some sort of kidney disease slash failure, but it does sound like if potassium levels are extremely important for a variety of people who have very specific health concerns,
00:07:09
something like this, it's perfect, and you share it with your doctor and that's, you know, where you go from there.
00:07:17
Yeah, I liked the idea in the TechCrunch article about this where somebody was like, you know, I have this kind of disease and I just have to guess whether it's okay to eat something that is high in potassium.
00:07:31
Sometimes I need to eat it, sometimes I shouldn't eat it, and this is the kind of thing that can make it easier for you.
00:07:36
So good stuff there.
00:07:38
Yeah.
00:07:39
Well, and, you know, I think that this, I know this is very specific, but it extends to, you know, all sorts of folks who are like, you know, sugar, right, you know, diabetic situations,
00:07:51
when we can know more about how our body is going to react to something that we're about to ingest, we are, we are all the wiser.
00:08:00
Yeah.
00:08:01
That's the interesting thing about this is we have lots of companies making tech for measuring glucose because of diabetes, but we just don't have that many products yet.
00:08:10
It seems like we have a good number of startups working on this, but for potassium.
00:08:14
So I'm actually curious, what other medical conditions folks in the audience think would be good to have companies working on?
00:08:22
There's probably a company working on it if you can think of it, but, you know, send us, send us your ideas for that feedback at dailytechnewshow.com.
00:08:30
Indeed.
00:08:31
Well, everyone, it is Christmas Eve, December 24th, and many of you probably have been reading year in reviews.
00:08:40
It's that time of year where everybody wants to do the year in review thing.
00:08:46
What was the best song or your favorite song?
00:08:48
What was, you know, the movie you liked the most?
00:08:50
The coolest car of 2024, best everything.
00:08:54
So here on DTNAS, thought we'd go around the horn and talk about what we liked most about tech, either something that we saw, something that we bought, something that we would like to buy in the future,
00:09:05
all of the things.
00:09:06
So this one, I'm going to go first.
00:09:09
So this one is not going to surprise anybody because I've been singing its phrases since I got it.
00:09:14
That is my M4 Pro Mac mini.
00:09:17
I got it as soon as it was, you know, the day it was announced, I ordered it.
00:09:22
It was somewhat of an issue to finally to procure it because I ended up being out of town when it was delivered and I was super worried that somebody was going to steal it off my porch.
00:09:34
Thankfully, that didn't happen.
00:09:36
But now that it, and it's powering my situation now, it is small and mighty and it is, I mean,
00:09:46
for a lot of things, you know, if you're like opening a website, it's like, wow, you know, the M4 Pro is not really going to wow you because it's just internet stuff.
00:09:56
But for a lot of, you know, processes that I do, for example, you know, encoding something or, you know, bouncing a video type thing,
00:10:07
I mean, it is just fastest heck.
00:10:09
I love it.
00:10:10
I am still very much a desktop person at heart.
00:10:14
I've got a laptop.
00:10:15
I've got an iPhone, but otherwise I like being in an office with a desktop situation.
00:10:20
And so the idea of like, what is a desktop now is, is very, very different than it used to be.
00:10:27
Yeah.
00:10:28
No, there was something you were waiting for all year long, right?
00:10:31
It was, that MacMoney to get announced, right?
00:10:35
You hopped on it.
00:10:36
I did.
00:10:37
I did.
00:10:38
Yeah.
00:10:39
My old MacMoney is behind us.
00:10:40
I'm sorry.
00:10:41
I still love you.
00:10:42
You're honored.
00:10:43
In an honored place up there on your shelf behind you.
00:10:46
Very much.
00:10:47
Yeah.
00:10:48
Like it's not even, you know, it's not, it's a crematorium at this point.
00:10:52
We'll figure out something.
00:10:53
But yeah, this is your MacMoney.
00:10:55
The new MacMoney and listen, you know, this is, you know, all with, with what I picked of the specs and, you know, shipping and tax is about $1,600,
00:11:06
you know, all in.
00:11:07
And that's a lot.
00:11:08
It's a lot of money.
00:11:09
And I, I feel like I, you know, it's a good purchase.
00:11:13
I'm, I'm very happy with it.
00:11:16
Brian, what about you?
00:11:17
Oh, man, this year I was challenged.
00:11:20
My wife said, Brian, I'm going to need you to pick something to want for Christmas and you're not going to get it until Christmas morning.
00:11:27
And I was like, well, that's novel.
00:11:28
Normally, I just buy whatever I want.
00:11:30
So in this case, I looked at my wrist and I'm currently still rocking my Garmin Vivo Active 4 from four and a half years ago.
00:11:38
But I wanted to poke in, see what the upgrades look like.
00:11:41
And it turns out that the cream of the crop on the Garmin side of things is the new Phoenix line, the Phoenix eights.
00:11:46
But then I was paralyzed because there's the motion and pixel kind of an e ink type display that is very battery conservative.
00:11:53
And then there's the OLED one as well.
00:11:56
But the, uh, the motion and pixel one, the, the mip display, it's got a solar charger on it.
00:12:02
So it goes, yeah, it goes, it goes like a month and a half.
00:12:06
If you don't use any of the GPS stuff, if you do the GPS stuff, you get three weeks on one chart.
00:12:11
So you get all.
00:12:12
That's incredible.
00:12:13
Oh, it's incredible.
00:12:14
And for somebody who loves all of those metrics about knowing whether or not I got my full sleep, having it take measurements at all times, we talked earlier about that gentle nudge of knowing whether or not you're living your life in the green zone.
00:12:27
This thing checks all of the boxes and keep in mind, I'm only going on based on the thousands of five star reviews.
00:12:34
And I remember being positively giddy reading the, uh, reading the, the, the, the pros, all of the pros listed listed and under cons, the only thing they could put on there was,
00:12:44
uh, not going to lie.
00:12:46
We wanted to ding it half a star based on how expensive it is, but truly it's worth it.
00:12:50
So, so I gave all the way in and so I'm going to be like a kid having a hard time falling asleep tonight because tomorrow I get my Phoenix eight.
00:12:59
So, so you are picking this as your tech of the year without having opened it yet.
00:13:03
Yes.
00:13:04
Uh, just like the Golden Globes nominated Squid games without it being premiered.
00:13:08
I was going to make the exact same comparison.
00:13:11
It's right.
00:13:12
I love that idea that you're like, I get a watch tomorrow, I can fill you up.
00:13:17
I'm just, you know, like, I hope it's in that square box, you're like, I know what it is.
00:13:22
Yeah.
00:13:23
Yeah.
00:13:24
Yeah.
00:13:25
I wouldn't put it past my wife to like prank me with a cold or something.
00:13:27
If there was that study that say sometimes spoilers raise your, uh, your involvement, raise, your experience.
00:13:34
This is an example of that, right?
00:13:36
You, you, you've been spoiled.
00:13:37
What's in the box, but it's still fun.
00:13:39
Oh, no, I'm having the appropriate level of agony having to wait to get my hands on this.
00:13:45
Yeah.
00:13:46
All right, Tom.
00:13:47
It's your turn.
00:13:48
What was your big tech of the year?
00:13:49
You know, I, I thought about this a lot.
00:13:50
Uh, and the thing I think has probably changed my daily life the most is the, uh, Logitech K 780 that I bought in August.
00:14:00
Uh, it is a multi device wireless keyboard.
00:14:04
Uh, you can do Bluetooth or, or you can do a, uh, RF, but I'm, I'm using it in Bluetooth mode.
00:14:09
Uh, and it's got really long battery life, uh, the, the reviews said it had really long battery life.
00:14:15
And I, I put the batteries in in August and have not changed them yet.
00:14:18
So it's still going, uh, I haven't really used the multi device part of it, but it's got three buttons at the top that let you switch between three different devices.
00:14:30
So I can have it.
00:14:31
And I do have it connected to my iPad, uh, my laptop and my phone, uh, and I used it once on an airplane to, uh, type on my phone, but I really only did it because I wanted to do it.
00:14:45
I was, there wasn't really a use case for, I was like, yeah, I don't know if I'm ever going to really do this where I pull it out.
00:14:52
Like I might as well just pull out my laptop at that point, uh, but you can do it.
00:14:55
And if you need to type something on the phone and your stuck, this is a great way to do it.
00:14:59
Um, my only complaint is that it lays very flat, uh, on the, on the table.
00:15:05
So I, I actually use a, a little piece of, of packing plastic to prop it up.
00:15:09
Um, but other than that, I, I love it.
00:15:12
It's, it's, it's got a wedge where you can sit a tablet or, or a phone in the top, uh, and use it that way, but I generally just use it with the laptop.
00:15:19
It connects really fast, uh, it's very quiet.
00:15:22
It's a very soft keyboard.
00:15:24
Uh, so if you have, as a mechanical keyboard person, how do you feel about that?
00:15:29
Uh, uh, it, it stops everyone from complaining about the sound of my typing.
00:15:33
So I like it, um, uh, complaints about my typing fell off, uh, quite a bit in August after I got this, um, what's now, what about those, uh,
00:15:43
they look like circular looking keys, dude?
00:15:46
Yeah, yeah.
00:15:47
So I'm really worried about my fingers sliding off to the side, but, but, but is everything feel natural?
00:15:53
I was curious about that too, but, but the key action is really good.
00:15:56
Uh, and so I, I forget that they're circular at this point and it just feels really nice and, and it is.
00:16:02
Uh, it's, uh, it's very smooth on F and J.
00:16:05
You're fine.
00:16:06
Oh, you know what?
00:16:07
Sarah mentioned how much the Mac mini is, uh, and I, I think I mentioned that the watch I wanted was, uh, like a little over a thousand dollars.
00:16:14
How much is what it is?
00:16:15
500.
00:16:18
Yeah.
00:16:19
This, uh, this sells for $55.
00:16:20
Ooh, stockings.
00:16:21
Yeah.
00:16:22
I mean, if you have a very big stockings, yeah.
00:16:24
Yeah.
00:16:25
Yeah.
00:16:26
But, you know, I mean, you could stick it in there.
00:16:27
Yeah.
00:16:28
You know, stick it wherever.
00:16:29
Yeah.
00:16:30
Put it in the tree.
00:16:31
We'll have a link in the show notes.
00:16:32
If you want to do that, uh, we'll, uh, Amazon.
00:16:33
Um, so yeah.
00:16:34
That, that's mine.
00:16:35
Uh, it's a keyboard, but it really practically has changed a lot.
00:16:39
So I like it.
00:16:40
For sure.
00:16:41
Yeah.
00:16:42
I mean, it's not the price.
00:16:43
It's the, it's the motion.
00:16:44
Right.
00:16:45
Uh, all right.
00:16:46
Roger, you've got to pick as well.
00:16:48
What do you want?
00:16:49
Yeah.
00:16:50
So I mentioned this, uh, a few weeks back, but I got this little tiny PC, which is essentially the guts of a laptop shoved into a, a little Mac mini sized, uh, PC case.
00:16:59
In this case, what I got, uh, was a 5800 AMD 5800H, uh, that has the, uh, uh, has a APU.
00:17:09
So it has a built-in, uh, integrated GPU that doesn't suck.
00:17:12
Uh, I paid 300 bucks for it came with 32 gigs of DDR4, 3200 gram and, uh, 500, or, uh, 512 gigabytes, uh, SSD.
00:17:21
I spent another 90 bucks to get, uh, one terabyte, uh, normal, two and a half inch SSD.
00:17:27
I got an enterprise drive.
00:17:29
So it has like triple the, the right capacity of, of your standard consumer one.
00:17:34
Uh, but all in, I'm running the show, I'm doing the show right now on it and I love it.
00:17:39
It's great.
00:17:40
It's, uh, it's small.
00:17:41
It's quiet and, uh, you know, if you need something that I, you just need a, a very low cost, but very functional PC, maybe not the best for gaming.
00:17:50
I played a few games last night.
00:17:52
Uh, it works for that, not, not to an amazing effect, uh, but for 300 bucks, uh, you could do worse.
00:17:58
Mm-hmm.
00:17:59
Excellent.
00:18:00
Uh, that, that's fantastic.
00:18:02
And, uh, folks, we'll have links to all four of these products in our show notes at dailytechnewshow.com.
00:18:08
So check that out if you have a question about this or anything else and you want to know our email address, here is me telling you your email address, not your email address, my email address.
00:18:18
It's feedback, feedback, uh, our email address, uh, does go to almost all of us feedback at dailytechnewshow.com.
00:18:30
When we talk about LLMs and generative AI models on the show, which we often do, 2024, has to be a record of some kind,
00:18:41
we focus on things like how it affects business or how government works.
00:18:47
And of course, what it does for us, the consumers.
00:18:50
So Brian, you are using LLMs every day and maybe in some creative ways that people might not know about.
00:18:57
Tell us a little bit about your, your strategy.
00:19:00
Yeah.
00:19:01
If we're going to package this up nicely for television, we call it three awesome LLM hacks with Brian.
00:19:06
Uh, the first one is, we're in this wonderful age where all of those soft skills, those language, uh, implying certain things without saying them or whatever.
00:19:16
Those traditionally have not been useful skills to have in programming or for internet searches.
00:19:22
You want to be very specific.
00:19:23
Now that whole world is flipped upside down and you could get astonishing answers from LLMs if you ask the right questions.
00:19:30
For example, uh, let's say you hypothetically have a YouTube channel and you need to manufacture a lot of music.
00:19:37
And what you want is to use audio to just create a sound alike for a temp track.
00:19:43
Let's say hypothetically you want, uh, the Beastie Boys groove homes to be in the background.
00:19:48
This kind of, uh, I don't know, I don't speak music words.
00:19:52
That's the problem, right?
00:19:53
How do you create a sound alike?
00:19:55
So you go to audio, you say, give me a sound alike track for groove homes and it says, I'm sorry, Dave.
00:20:01
I can't do that.
00:20:02
I would never violate copyright.
00:20:04
And so you can't be.
00:20:05
And so either a, I could learn a whole bunch of music talk things to say, I don't know what instruments are in that particular track or how to describe.
00:20:13
I don't know the difference between upbeat downbeat or the daytime or nighttime funk.
00:20:18
I don't know none of that.
00:20:19
So I go over to chat GPT and I'm like, hey, how would you describe groove homes by the Beastie Boys?
00:20:27
And I was like, if you wanted to create that song, what prompt would you use?
00:20:31
And it would give a whole bunch of like, well, I would just say give me a temp track.
00:20:34
I was like, yeah, but can you pull that off without mentioning the track that you're copying?
00:20:40
And then it says, you bet it boss.
00:20:41
And then I just copy the one and paste it on the other surprise.
00:20:44
Now I have a temp track that I like very, very much that I own the rights to because whenever I license music, the company gets bought down the road.
00:20:52
And then I get demonetized as a big pain in the butt.
00:20:55
So get, turn the AIs against each other is my first bit of advice.
00:20:59
Don't tell it that you're, that you're subverting something, but just find out how it would write the prompt for things and then copy and paste it.
00:21:09
The second tip I have for everyone is that most jobs have a secret language.
00:21:15
And think about your day job, whatever it is you do.
00:21:18
There are certain things people can ask that instantly you know that they're an equal or a player of even if it's in a peripherally related way.
00:21:26
For example, when I toured for 20 years doing college magic shows, I was always flying Southwest Airlines.
00:21:32
So I knew to immediately go to the back of the plane and I had about two minutes to strike up a conversation asking, oh, hey, AM shift isn't it great?
00:21:42
Everybody's well behaved, right?
00:21:43
And then if it was later in the day, I'd say PM shift, everybody's having a good time on these, right?
00:21:48
And then where are you based out of?
00:21:51
And then all of these inside jargonny things, is that really jargon, or is that just being nice to a, well, you also, but the fact that you know that what the two shifts are and then,
00:22:06
yeah, it's the words you choose, right?
00:22:09
Yeah.
00:22:10
Exactly.
00:22:11
And so those secret light.
00:22:12
And so, you know, what happens is that they would immediately ask like, oh, do you fly with us?
00:22:15
It's like, nope, and then that gives me the opportunity to casually mention the fact that I'm on the road 300 days out of the year and fly Southwest Airlines exclusively.
00:22:25
And miraculously, I would always get drinks ahead of everyone else and they would never charge me.
00:22:31
But you can do that exact same thing.
00:22:34
If you write some prompts, start with your field of expertise and see if you can get the chat.
00:22:40
That way is a stage hypnosis techniques of finding a concrete foundation to build your relationship on, get it to a place where it is throwing jargon that you personally know is accurate and then freeze that prompt,
00:22:52
freeze that session, copy, paste it as if you're sharing it to someone else, then your next question can be, hey, I'm about to call and negotiate down my rate on a credit card.
00:23:04
What have you?
00:23:05
What are some inside jargon that will be meaningful to somebody working the phone queue at Chase or I'm filing late for this property tax thing?
00:23:14
What is some inside jargon that I could drop that will get their attention?
00:23:18
And the prompt will understand what you're going for and it'll just give you a bunch of keywords to drop in there.
00:23:24
And then you can approach it with total confidence and sort of mentally skip the line.
00:23:29
And then the last thing is something that I just started this week is I announced, you know, here at Modern Rogue World Headquarters, we got a bunch of buildings, we got a bunch of different systems and unfortunately all of it is locked inside my dumb stupid head and I have to remember what things need to play well with what devices.
00:23:49
And so I told the promo chat GPT, hey, I want to map out everything and it says great.
00:23:55
Well, and it gave me a giant bulleted list of like 10 different systems and it told me to write down all of the things and I was like, no, no, no, no, no, I don't want to do all of this.
00:24:05
So I do it, ask me one question at a time and I'll answer each one question and you build the whole thing.
00:24:12
That's great.
00:24:13
And so first it's, you got any Sono speakers?
00:24:16
Yes, I got them here, here, here and here.
00:24:18
And it's like, what about Z-Wave devices?
00:24:20
I'm like, yes, I have this type of device here, here, here and here.
00:24:25
And just by doing it one at a time, I'm going to keep doing this.
00:24:28
If I do five a day a month from now, I'm going to have an extraordinarily detailed document and I'll be able to say, hey, where are the inefficiencies?
00:24:37
Where are this?
00:24:38
How do I, what devices should I have that normally in this environment could whatever?
00:24:43
And then you could say, hey, write me a script so that when a new guest arrives, I can plug in one thing and it'll automatically change all the codes and update all across campus and set permissions and day and day and day and day and but it all begins with setting up that simple,
00:24:57
asking one question at a time in a persistent thread.
00:25:01
Yeah, it really shows the need to understand how to use the prompts, right?
00:25:09
It's like, these tools are not yet user-friendly in the sense of you don't need to think.
00:25:14
You need to think a little bit and then they can become very user-friendly, but you have to tell them to be user-friendly, I guess.
00:25:19
And you also have to draw boundaries like no or not good enough.
00:25:25
These are phrases that unless you're a sociopath working with a team, most of us don't think to say or do those, but this is a device that has no feelings and you must talk to it as though your Bill Murray enscruged and once you get over that,
00:25:42
oh.
00:25:42
Which is, I think that's hard for a lot of people, you know, I still say, please do this for me.
00:25:48
Sure.
00:25:49
You know, because I'm like, I don't want to be rude, but it's not a person.
00:25:53
And in fact, if you use too much soft language, it will respond the way human would tend to, which is continue to roll over on you, whereas if you flatly say,
00:26:05
no, that's unacceptable, you can't do that, I need it as this.
00:26:09
Like for example, the one question thing was a game-changer, I'm like, here's how it's going to go.
00:26:15
You're going to keep asking one question at a time forever and then eventually you're going to get all the answers go.
00:26:19
And it asked me a question that had me like, I don't know the particular model number of the router I have and I can't check it right now.
00:26:27
So I just responded with, asking that later, I can't get it for you.
00:26:31
What else you got?
00:26:32
And eventually we'll get there, but for now, it's like, no, keep going.
00:26:38
That's really really cool.
00:26:39
This is good stuff and we'll have like short notes about this in the show notes.
00:26:45
But if you were to tell people how to figure out how to do this on their own, is there resources or things that people can go to?
00:26:53
And this is the exciting thing about it being the dawn of an event, nobody, not even the people creating this stuff, know what it's going to be used for.
00:27:01
And so here's a tidbit, no matter what your field of expertise is, think about this.
00:27:07
They're almost certainly is not a world's foremost expert in AI within your subcategory.
00:27:15
You today can become the world's foremost expert in your category, whatever it is, the AI expert of plumbing, the AI expert of magic or whatever.
00:27:28
All of those are open slots and you can take one.
00:27:31
All you have to do is start a conversation.
00:27:33
If you don't know how to start, literally tell chat GPT I would like to become the world's foremost expert in blank, an AI blank and it will come up with a plan that you just follow.
00:27:44
That's amazing.
00:27:45
All right, before we get out of here, let's check out the mail bag.
00:27:49
We got a really good one from Michael and an audio clip.
00:27:53
Michael had some thoughts about the accessibility benefits that Rayban by Meta provides and may provide in the future, especially the Be My Eyes hands-free integration.
00:28:03
Let's hear from him.
00:28:04
In response to Monday's Daily Tech News Show episode related to Meta Raybans, I wanted to share some advantages that these give blind or sight impaired individuals.
00:28:16
There's of course the look and describe this feature, which if you didn't know, you can tell Meta to look and describe things and it will do an okay job at describing them.
00:28:26
But more importantly, I feel, is earlier this year, Be My Eyes and Meta announced a partnership where you can say, "Hey Meta, Be My Eyes," and all of a sudden you're connected with a volunteer who can also see through that camera.
00:28:40
Now, that camera I've learned is not the best for reading text, especially smaller text, but it's an amazing opportunity to quickly get someone to assist you hands-free without having to pull out your phone.
00:28:54
And there are paid services out there, like Ira, that have a similar experience, not quite as smooth, but something to consider when exploring the Meta Raybans.
00:29:04
Meta, hope you're listening.
00:29:07
Please, don't add too much dependency on this alleged screen to come to Meta Raybans next.
00:29:14
That was great.
00:29:15
We were asking about accessibility uses, and that was wonderful for Michael to respond with that.
00:29:22
Thank you, Michael.
00:29:23
Indeed, thank you, Michael, and thank you to everybody who writes in audio notes.
00:29:29
I mean, we love them, but you can write us about anything.
00:29:33
And feedback at dailytechnewshow.com.
00:29:35
Brian, Rushwoods, you better be on the show a little bit more often.
00:29:39
Okay.
00:29:40
That'll be my New Year's resolution.
00:29:41
His many more visits, so I'll come bearing more gifts.
00:29:44
Indeed.
00:29:45
I know we know you're all very busy, and you're all over the place.
00:29:53
So let folks know what you're up to lately and where they can keep up with you in the New Year.
00:29:58
I'll tell you what.
00:29:59
What you could do is just search for the words world's greatest con or head on over to world's greatest con.com.
00:30:05
We just dropped a new two-hitter, two episodes back-to-back.
00:30:08
If you remember crazy Eddie, those wacky commercials from the early 1980s, what you may not remember is that the dude was running a mountain of grift.
00:30:17
It was incredible.
00:30:18
He ended up an international fugitive and stole millions of dollars.
00:30:23
Turns out you shouldn't do an IPO when you're building his house of cards.
00:30:28
That's good advice.
00:30:30
Patrons, stick around for the extended show, Good Day Internet.
00:30:33
Our producer, Amos, is going to talk about how he used chat GPT to plan a five-day trip to move his wife's things from down near Seattle to Wastilla,
00:30:43
Alaska.
00:30:44
How did Chad do as his assistant, Amos, is going to tell us.
00:30:49
This is our last live show of 2024, and our last show, our last live show.
00:30:54
We will be back with live shows Thursday, January 2nd, 4 p.m.
00:30:58
Eastern, 100 UTC as usual, and you can find out more at dailytechnewshow.com/live.
00:31:06
But worry not, our holiday show schedule is pretty packed.
00:31:10
It starts the Thursday, December 26th, with our annual listener show It's a good one this year, with a lot more fun episodes until we see you next year.
00:31:18
Happy Holidays.
00:31:22
With a DTNS family of podcasts, helping each other understand.
00:31:29
- All right, this problem.
00:31:31
(laughs)
00:31:32
