ChatGPT - This is pretty awesome :)

[joe.pistell]

anyone using ChatGPT to help with analyzing paid search efforts?

RB, In my AI use (2+ hrs a day in AI), AI doesn't do math well.

 

[Carsten]

chatGPT makes stuff up.

I'm starting to look at just rolling something on my laptop like: WebLLM | Home
There are other models out there.

However, AWS (amazon) just announced they are upgrading their Q and some other things.

I am thinking we are still 1-2 years away from AI that is truly useful.

 

[joe.pistell]

I am thinking we are still 1-2 years away from AI that is truly useful

 

[Rick Buffkin]

Let’s analyze Uncle Joe's AI experience using a formula that quantifies his usage over time and attempt to assess AI’s knowledge of mathematics. Here's how we can approach this:

---

Formula for AI Experience Level​

We define AI Experience Level (E) as a function of time spent daily (T) and total duration in days (D):

E=T⋅DE = T \cdot DE=T⋅D
Where:

• TTT: Time spent daily using AI (in hours, T=2T = 2T=2 for Uncle Joe).
• DDD: Total number of days of usage over the past 6 months (D=6⋅30=180D = 6 \cdot 30 = 180D=6⋅30=180 days).

Formula for AI Knowledge Level at Math​

For simplicity, we assume a relative score, KKK, reflecting AI’s perceived knowledge level in mathematics. This could range from 1 to 100, representing Uncle Joe's subjective rating over time. Let’s hypothesize KKK as proportional to cumulative AI usage experience (E), but capped at a maximum level, Kmax=100K_{\text{max}} = 100Kmax=100:

K=min⁡(Kmax,α⋅E)K = \min(K_{\text{max}}, \alpha \cdot E)K=min(Kmax,α⋅E)
Where:

• Kmax=100K_{\text{max}} = 100Kmax=100: Max score for AI's math knowledge.
• α\alphaα: Scaling factor (α=0.01\alpha = 0.01α=0.01 to keep KKK realistic).

---

Graph Representation​

We’ll graph:

1. Experience Level (E) over 6 months (days: 1 to 180).
2. Knowledge Level (K) over the same period.

Let me compute and plot this.



The graphs above represent:

1. Uncle Joe's AI Experience Level (E): This increases linearly over time because he spends a constant amount of 2 hours daily using AI.
2. AI's Knowledge Level at Math (K): This starts growing proportionally with the experience level but caps at a maximum level of 100 due to the imposed limit on AI's perceived maximum capability.

Insights:​

• Uncle Joe's experience grows unbounded as long as he continues using AI.
• The perceived knowledge level of AI in math saturates at a certain point, reflecting the hypothesis of diminishing returns with extended use.

If you'd like to explore further adjustments to this model or focus on a specific scenario, let me know! Keep up the curiosity!

 

[Carsten]

Did you just suggest that Uncle Joe is the Ollama?



Note: AWS is massively doubling down on Ai and have even beefed up their server farms. Still over all it just seems they are trying to draw you into their data lake storage systems and using their Q product to display your data.

Ai is useful but it still is rogue element that needs to be fact checked.

 

[joe.pistell]

Uncle Joe sees the world differently...
With self publishing being the cornerstone of SMB's, here's a 1st pass at AI creating a:

LinkedIn Professional Image Assessment​

Professional Background​

Technology CEO/Founder in automotive software (FRIKINtech) and digital community leadership (DealerRefresh). Strong focus on improving dealer-customer relationships through technology solutions.

Assessment Purpose​

Enhance executive presence to build trust with automotive dealers and technology partners.

Overall Score: 72/100​

Industry Alignment (28/40)

• Attire: 12/20 (Polo lacks executive presence)
• Professional polish: 7/10
• Industry styling: 9/10

Technical Quality (29/40)

• Lighting: 10/15 (Overhead shadows)
• Camera/framing: 12/15
• Background: 7/10 (Car interior lacks distinction)

Brand Consistency (15/20)

• Role alignment: 7/10
• Profile cohesion: 8/10

Critical Priority Changes​

1. Professional Wardrobe Update

• Brooks Brothers Sport Coat ($498)
• Custom dress shirts, 3-pack ($285)
• Timeline: 1 week

1. Professional Headshot Session

• Executive package with top photographer ($650)
• Indoor/outdoor options
• Timeline: 2 weeks

1. Lighting Equipment

• Ring light with stand ($89)
• Portable LED panel ($129)
• Timeline: Immediate

Implementation Timeline​

Immediate

• Order lighting equipment
• Schedule photographer
• Book wardrobe consultation

Short-term

• Complete wardrobe updates
• Professional photo session
• Update all social platforms

Long-term

• Quarterly wardrobe refresh
• Bi-annual photo updates

Budget Overview​

Essential: $1,650

• Professional photography
• Basic lighting
• Core wardrobe pieces

Full Enhancement: $2,500

• Complete wardrobe refresh
• Premium lighting setup
• Quarterly touch-up sessions

Hey Alex. The good news is your 72 score was wayyy better than @todd.smith's 59

 

[joe.pistell]

Uncle Joe sees the world differently...

ARE YOUR ADS OEM COMPLIANT?
I fed AI the Chevy Ad standards pdf, then I asked it if our ad was compliant... ouch.

Several compliance issues with this dealer ad:
Typography & Fonts:

• Not using required Chevy Sans typeface
• Incorrect font styling for price point and lease terms
• Non-compliant number formatting

Colors:

• Red background isn't in approved Chevrolet color palette
• Should use Deep Gray (#393839) or approved vehicle-related colors

Branding:

• Dealer logo not following co-branding lockup guidelines
• Missing required Together let's drive™ tagline
• Missing proper EV brandmark treatment

Layout:

• No rhombus element where appropriate
• Improper spacing around dealer branding
• Disclaimer text not in correct Chevy Sans Narrow font

Recommendations:

1. Switch to Chevy Sans font family
2. Use approved color palette matching vehicle color
3. Implement proper EV branding lockup
4. Add rhombus element with correct 63° angle
5. Reformat dealer logo according to co-branding guidelines
6. Use proper typography hierarchy for offer details
7. Follow standard disclaimer formatting

 

[jscole86]

ARE YOUR ADS OEM COMPLIANT?
I fed AI the Chevy Ad standards pdf, then I asked it if our ad was compliant...

That's a great use for it!

 

[Carsten]

I'm not gonna be getting chatGPT pro if what I'm hearing is true.
I'd rather buy myself a mac mini and do it locally.

 

[joe.pistell]

I am in AI 2-3hrs a day. I have subs to Meta, GPT, Claude, Gemini, Perplexity. Why?

One Example:
I am building a merchandising system for @BillKVMotorCo and Bill's photo process is so excellent, that we're created a new layout to complement his team's work (kudos to @Will G. !) I am pressing all build data into Claude 3.5 for data normalization and categorizing and attribute weighting too. It's not cheap but the output is excellent. This was not possible 6 months ago.