Artificial Intelligence: How are you using it in everyday life?

ISUKyro

Well-Known Member
Oct 28, 2006
13,124
2,706
113
Houston, TX
I use it a lot in my job. Some of the ways I have used it:

- Take text I wrote and make it more “marketing-y” or concise
- Check an Excel VBA macro I wrote for errors, and make improvements to it - the problem is sometimes it generates VBA code of its own that doesn’t work, so it takes a round two to ask what it did wrong
- Create a podcast of text that is too boring to read (NotebookLM does this well)
- Summarize meeting notes
- Get me an initial draft of a work email that I am having trouble getting started. From there, I improve it on my own and fact check
- After I pull any intellectual data or information out of text, it can either help me make something more or less technical to translate it to the appropriate audience
Excel is where I use it the most as well. I’m ok with Excel, I’m just good enough to write a prompt to AI and make t has gotten me some great formulas and VB code that I otherwise would not thought of as quickly
 
  • Friendly
Reactions: Angie

dmclone

Well-Known Member
Oct 20, 2006
21,548
5,895
113
50131
I've been using Gemini and ChatGPT this whole time. Yesterday, I tried using Grok/XAI and it seems to provide really good answers but they are REALLY long.
 

dmclone

Well-Known Member
Oct 20, 2006
21,548
5,895
113
50131
Here is an example. Only ChatGPT was able to figure this one out:

When this script is run, it always says the same thing twice in a row. Can you add some kind of condition or delay to stop this from happening. metadata: name: Someone is at the door description: Script someone is at the door automations: - starters: - type: device.event.PersonDetection device: Front Door Camera - Front Door actions: - type: assistant.command.Broadcast devices: - Google kitchen - Kitchen message: There is a person at the front door

ChatGPT said:​

It looks like you're using a script for a home automation system, where the action is triggered when someone is detected by your front door camera. The reason why the message might be sent twice could be because the event is being triggered multiple times or too quickly. To prevent this, you can add a condition or a delay to ensure that the action only happens once within a certain time frame.
You can modify the automation to include a delay or a condition to check if the event has already been triggered recently.

Solution 1: Add a delay to prevent multiple triggers in quick succession​

Add a delay in the actions to ensure the message only broadcasts once, even if the event is detected more than once within a short time.
yaml
Copy
metadata:
name: Someone is at the door
description: Script someone is at the door
automations:
- starters:
- type: device.event.PersonDetection
device: Front Door Camera - Front Door
actions:
- type: assistant.command.Broadcast
devices:
- Google kitchen - Kitchen
message: There is a person at the front door
- type: delay
duration: '00:01:00' # Delay of 1 minute before the next event can trigger

Solution 2: Use a condition to check if the event has already been triggered recently​

You can also use a condition that checks if the event has occurred within a set time period, and if it has, it won’t trigger again until the time period has passed.
yaml
Copy
metadata:
name: Someone is at the door
description: Script someone is at the door
automations:
- starters:
- type: device.event.PersonDetection
device: Front Door Camera - Front Door
actions:
- type: condition
condition:
type: time
after: '00:01:00' # Only allow the action if 1 minute has passed
- type: assistant.command.Broadcast
devices:
- Google kitchen - Kitchen
message: There is a person at the front door

Solution 3: Use a presence sensor or variable​

Alternatively, you could use a presence sensor or a boolean variable to track whether the event has already been triggered and prevent multiple messages. This approach may require more configuration depending on the platform you're using.
Let me know if you need help with this!

You said:​

[delay] is an unknown type name. Expected types: [device.command, device.command.ArmDisarm, device.command.BrightnessAbsolute, device.command.Dock, device.command.SetInput, device.command.NextInput, device.command.PreviousInput, device.command.StopLightEffect, device.command.LightEffectSleep, device.command.LightEffectWake, device.command.LightEffectPulse, device.command.LightEffectColorLoop, device.command.LockUnlock, device.command.OnOff, device.command.OpenClose, device.command.StartStop, device.command.PauseUnpause, device.command.ThermostatTemperatureSetpoint, device.command.ThermostatTemperatureSetRange, device.command.ThermostatSetMode, device.command.Mute, device.command.SetVolume, device.command.ColorAbsolute, device.command.AppInstall, device.command.AppSearch, device.command.AppSelect, device.command.BrightnessRelative, device.command.SelectChannel, device.command.RelativeChannel, device.command.ReturnChannel, device.command.Cook, device.command.Dispense, device.command.Charge, device.command.SetFanSpeed, device.command.SetFanSpeedRelative, device.command.ReverseFan, device.command.Fill, device.command.SetHumidity, device.command.HumidityRelative, device.command.FindMyDevice, device.command.Reboot, device.command.RotateAbsolute, device.command.ActivateScene, device.command.TimerStart, device.command.TimerAdjust, device.command.TimerPause, device.command.TimerResume, device.command.TimerCancel, device.command.MediaStop, device.command.MediaNext, device.command.MediaPrevious, device.command.MediaPause, device.command.MediaResume, device.command.MediaShuffle, assistant.command.OkGoogle, assistant.command.Broadcast, time.delay, home.command.Notification, etc].

ChatGPT said:​

It seems like the automation system you're using doesn't recognize the delay type as a valid action type, but it does recognize time.delay. Let's adjust the automation to use time.delay instead of delay. Here's the corrected version of your script:

Corrected Version with time.delay:​

yaml
Copy
metadata:
name: Someone is at the door
description: Script someone is at the door
automations:
- starters:
- type: device.event.PersonDetection
device: Front Door Camera - Front Door
actions:
- type: assistant.command.Broadcast
devices:
- Google kitchen - Kitchen
message: There is a person at the front door
- type: time.delay
duration: '00:01:00' # Delay of 1 minute before the next event can trigger
In this version, I've replaced delay with time.delay, which should be recognized by your system.
Let me know if this works or if you need further adjustments!

You said:​

When I use this, I get one error on line 16. There error is "Field [duration] is unexpected in the Struct." Here is the script: metadata: name: Someone is at the door description: Script someone is at the door automations: - starters: - type: device.event.PersonDetection device: Front Door Camera - Front Door actions: - type: assistant.command.Broadcast devices: - Google kitchen - Kitchen message: There is a person at the front door - type: time.delay # Adjust duration as needed. Supported units: sec, min, hour for: 5min duration: "00:01:00" # Delay of 1 minute before the next event can trigger

ChatGPT said:​

It seems that the field duration is causing the error because it isn't expected when you're using time.delay. In your automation system, it appears you can only use the for keyword to specify the duration for delays.
Here's the corrected script with for instead of duration:

Corrected Script:​

yaml
Copy
metadata:
name: Someone is at the door
description: Script someone is at the door
automations:
- starters:
- type: device.event.PersonDetection
device: Front Door Camera - Front Door
actions:
- type: assistant.command.Broadcast
devices:
- Google kitchen - Kitchen
message: There is a person at the front door
- type: time.delay
for: 5min # Delay for 5 minutes before the next event can trigger

Explanation:​

  • The correct way to define the delay duration in your system is by using for, followed by the desired duration (like 5min).
  • The line with duration: "00:01:00" was causing an error because it wasn't a valid field for the time.delay action.
Let me know if this works!

You said:​

That worked!

ChatGPT said:​

Awesome! I'm glad it worked!
If you have any more questions or need further adjustments, feel free to ask. Happy automating!
 

Angie

Tugboats and arson.
Staff member
SuperFanatic
SuperFanatic T2
Mar 27, 2006
28,585
13,579
113
IA
Excel is where I use it the most as well. I’m ok with Excel, I’m just good enough to write a prompt to AI and make t has gotten me some great formulas and VB code that I otherwise would not thought of as quickly
It really is great with it. I've been able to ask it questions when I was thinking my way around macros, and it's good at being a sounding board, then giving suggestions for tweaks.
 

Jer

CF Founder, Creator
Feb 28, 2006
23,580
23,427
10,030
I've been using it to summarize or put my stuff into layman's terms. For instance, I put all my recipes through it and asked it to create consistently formatted and measured step-by-step instructions. Helps the wife immensely learn how to cook for me.

I've also had some formal request letters and things where I don't necessarily know the best technical medical jargon but know what I need to communicate. Telling it how to take my inputs and put into a specific type of form letter, appeal, claim, etc has been great.
 
Last edited:
  • Like
Reactions: dmclone

CivEFootball

Well-Known Member
Sep 16, 2010
631
503
93
Used it to write the office wide invitation to tournament bracket. Then had it write a poem reminder to fill out the bracket which is delay delivering no sonner than tomorrow at 8am.
 

Cyclones01

Well-Known Member
SuperFanatic
SuperFanatic T2
May 23, 2008
2,347
1,199
113
Urbandale
I'm a software engineer and I use it every day. There are some tasks it can handle for me entirely and there are some things that I have learned not to even attempt to use it for. Either way, I still have to keep a close eye on it and make sure it's doing what I want. It doesn't come close to doing my job for me, but in aggregate, it allows me to move significantly faster.

It's like having a more junior engineer helping me do my work; they're not always correct but they are extremely fast. So even when they're wrong, we move past it quickly.
 
  • Like
Reactions: dmclone

BobTheHawkHater

Well-Known Member
Jan 21, 2008
161
388
63
I'm a software engineer and I use it every day. There are some tasks it can handle for me entirely and there are some things that I have learned not to even attempt to use it for. Either way, I still have to keep a close eye on it and make sure it's doing what I want. It doesn't come close to doing my job for me, but in aggregate, it allows me to move significantly faster.

It's like having a more junior engineer helping me do my work; they're not always correct but they are extremely fast. So even when they're wrong, we move past it quickly.
I've asked it to write bash scripts/functions to do stuff a few times and they didn't work but I got a lot of good ideas and I was able to fairly easily modify them to get something working quickly.

I write embedded software in C/C++, which means digging through tons of chip datasheets and application notes to figure out how to get some hardware to do something I've never worked on before. I can ask it simple "how would I ..." questions for a specific microcontroller and I'm just completely amazed at how it is able to come up with really useful high level info to educate myself before I dig into coding. After I've consumed that much, I can ask for additional info all the way down to writing some example code for me and it gives me a decent start. I can't even imagine where this will go in 25, 50, 100 years from now.
 

dmclone

Well-Known Member
Oct 20, 2006
21,548
5,895
113
50131
Thinking about asking it to fill out my bracket
I tried that this morning. Not so much my bracket, but what characteristics the winning teams have had in the last 15 years. It mentioned some things like strong guards, good defense, and how 3 point shooting has become more important. Of course it couldn't tell me which teams were strong in these areas.
 

KnappShack

Well-Known Member
May 26, 2008
23,825
32,148
113
Parts Unknown
I tried that this morning. Not so much my bracket, but what characteristics the winning teams have had in the last 15 years. It mentioned some things like strong guards, good defense, and how 3 point shooting has become more important. Of course it couldn't tell me which teams were strong in these areas.

Try Grok.

It gave me Duke, Auburn, Houston, UConn, Gonzaga, and Alabama as true contenders based on metrics that mirrored prior champions.

Some decent color around the choices too.

I don't follow basketball. These seem legit?
 
  • Like
Reactions: dmclone

dmclone

Well-Known Member
Oct 20, 2006
21,548
5,895
113
50131
Try Grok.

It gave me Duke, Auburn, Houston, UConn, Gonzaga, and Alabama as true contenders based on metrics that mirrored prior champions.

Some decent color around the choices too.

I don't follow basketball. These seem legit?
Well the high seeds normally do win the overall championship, so that's probably as good as anyone else's guess. Gonzaga is the only one that sticks out. Maybe they are going on past performance.
 

dmclone

Well-Known Member
Oct 20, 2006
21,548
5,895
113
50131
Damn, Grok does seem to do a good job. I asked who had the strongest backcourt this year and it spit out a lot of good info.
 

dmclone

Well-Known Member
Oct 20, 2006
21,548
5,895
113
50131
"Which teams are most likely to cover the spread". Wow again

OK, as I looked deeper it got some of the matchups wrong. it had ISU playing Samford. This has been my typical experience with AI. Don't trust and verify.
 
Last edited:

KnappShack

Well-Known Member
May 26, 2008
23,825
32,148
113
Parts Unknown
Well the high seeds normally do win the overall championship, so that's probably as good as anyone else's guess. Gonzaga is the only one that sticks out. Maybe they are going on past performance.

Here's it's write-up:

Why They Match Up: Gonzaga’s pick-and-roll offense ranks in the top 10 for efficiency, led by Ryan Nembhard and Graham Ike. They’re in the 90th percentile for pick-and-roll efficiency in 14 of the last 16 years, per Synergy, and their veteran roster (one of the oldest in Division I) mirrors champions like Villanova (2016, 2018). They also limit neutral-court losses
 

KnappShack

Well-Known Member
May 26, 2008
23,825
32,148
113
Parts Unknown
Here's it's thoughts about the good guys

Realistically, Iowa State’s title chances hinge on Lipsey’s health, avoiding off nights against mid-tier teams, and peaking against top seeds. They’re not favorites, but their metrics and style give them a puncher’s chance—think 10-15% probability, higher if they catch fire. So, yes, they can take the title, but they’re a long shot who’d need to play near-perfect basketball.
 
  • Like
  • Informative
Reactions: NWICY and SayMyName

gypsyroad

Active Member
Oct 24, 2023
81
176
33
I'm writing some Azure AI code in a language that isn't one of the MS docs typical selections like python or JS. I can take the provided http calls, drop it in ChatGPT and say "convert this to my language". I have to make some modifications but it generally does a good job. Significantly faster than doing it for each call I need to make. It often gives me options I didn't think of.