- AI with Kyle
- Posts
- AI with Kyle Daily Update 029
AI with Kyle Daily Update 029
Today in AI: Billionaires Bickering + GPT-5 Rollbacks continue
Kyle Balmer
August 14, 2025
Live Webinar: Learn how you can become an AI Trainer… get paid $1-$4k per hour …without needing a PHD in AI, or being a TED Talk level public speaker… 📅 TODAY ⏰ Check link for timezone 👉️ Register here
The skinny on what's happening in AI - straight from the previous live session:
Highlights
🧠 GPT-5 Gets Major Personality Update
OpenAI has announced significant changes to GPT-5, including the return of GPT-4 and new personality controls. Users can now choose between auto, fast, and thinking modes, with rate limits set at 3,000 messages per week for GPT-5 thinking mode.
The changes come after widespread criticism of the unified model approach, with users experiencing inconsistent routing between different quality tiers of GPT-5 (high, medium, low, and minimal). OpenAI is also working on updating GPT-5's personality to strike a balance between the warmth of GPT-4 and the more direct approach of GPT-5.
Here’s the full update.
Updates to ChatGPT:
You can now choose between “Auto”, “Fast”, and “Thinking” for GPT-5. Most users will want Auto, but the additional control will be useful for some people.
Rate limits are now 3,000 messages/week with GPT-5 Thinking, and then extra capacity on GPT-5 Thinking
— Sam Altman (@sama)
1:19 AM • Aug 13, 2025
Kyle's take: This is clearly damage control from OpenAI. They've realised that suddenly removing GPT-4, which many users had formed emotional connections with, was a mistake.
The new personality options and return of legacy models show they're trying to appease both technical users who want intelligence and those seeking emotional support. However, finding the right personality balance for GPT-5 will be challenging (to say the least!) - they risk satisfying neither group. Expect more friction!
⚔️ Billionaire AI Drama: Musk vs Altman
Elon Musk has threatened legal action against Apple over Grok's fifth-place ranking in the App Store, claiming antitrust violations while ChatGPT maintains the top position. Sam Altman responded by pointing out Musk's own manipulation of X's algorithm.
Musk responded maturely:
You got 3M views on your bullshit post, you liar, far more than I’ve received on many of mine, despite me having 50 times your follower count!
— Elon Musk (@elonmusk)
9:26 AM • Aug 12, 2025
Kyle's take: Hilariously Grok stepped in and told Musk that he was in the wrong and Altman was correct. Then GPT-5’s own account hopped in and called Grok a “good bot”…
Utterly bizarre.
This appears to be Musk seeking attention while GPT-5 dominates the news cycle. Fifth place for a new AI app is actually really impressive, but Musk's reaction reveals deeper concerns about Grok's adoption rates compared to ChatGPT's dominance. The irony of their own AIs mediating their argument is perhaps the most '2025' aspect of this entire situation.
More worrying is that these two men are spearheading our drive towards AGI and ASI…
💸 Perplexity's Surprising Chrome Bid
Perplexity AI has made an unexpected $34 billion offer to purchase Google Chrome, despite being valued at only $18 billion themselves. The bid comes amid antitrust pressure on Google to potentially divest parts of its business.
Kyle's take: This appears to be primarily a marketing move by Perplexity to draw attention to their own browser, Comet.
With Chrome's 3.5 billion users versus Perplexity's relative obscurity (15m browser users supposedly), the bid has successfully generated headlines and awareness for their competing product, even though there's little chance of the purchase actually happening.
And in the tiny chance it went ahead? Well, I’m sure Musk or Altman would hop in to help with the raise and snap up Perplexity themselves…with Chrome’s 3.5B users attached.
Member Question: How should we approach emotional attachment to AI models?
This is one of the most critical questions we face as AI becomes more sophisticated.
I don’t think using AI for companionship is inherently problematic - it's not dissimilar from journaling for instance - we need to be cautious about dependency.
The real challenge is drawing the line between healthy use and potentially harmful attachment, especially for vulnerable individuals. Who draws that line? Sam Altman? Elon Musk??? Hmm.
Rather than dismissing these attachments, we need to develop frameworks for responsible AI companionship while ensuring proper human support systems remain primary. That’s going to be a big project for humanity over the next couple of decades.
Want the full unfiltered discussion? Join me tomorrow for the daily AI news live stream where we dig into the stories and you can ask questions directly.