Published on LinkedIn and AmitabhApte.com on August 10, 2025
In spotlight this week: GPT-5 lands but not everyone’s cheering
The AI world has been holding its breath for GPT-5, the long-promised leap forward. Now it’s here. But instead of unanimous applause, the launch has landed like a blockbuster film breaking box office records while dividing critics.
OpenAI calls GPT-5 its most capable, reliable, and safe model yet, a multimodal workhorse for coding, writing, health, and complex reasoning. It’s faster, hallucinates less, remembers more, and can now work seamlessly across text, images, and code. Microsoft Copilot is already running on it, meaning millions will soon be using GPT-5 without even knowing it.
On paper, this is the AI assistant we’ve been promised:
- Longer memory & context so it can finally act like a long-term colleague, not a one-off chatbot.
- Multimodal fluency for integrated text, image, and code workflows.
- Enterprise-grade reliability & safety for regulated industries and mission-critical work.
My early take? This is a strategic reset, simplifying model choices for users while pushing benchmark-beating features that play well in health, enterprise, and developer spaces. But some of the most enticing tools, like Google Calendar integration, sit behind the pricier Pro tier, risking a fragmented user experience.
And the user feedback? A mixed bag. Some love the speed and precision. Others miss GPT-4o’s personality describing GPT-5 as shorter, blunter, and less emotionally intelligent. My bet: early quirks will be ironed out. Whether GPT-5 is better for day-to-day use than GPT-4 or GPT-4o will be decided not by benchmarks, but by how it feels in the hands of real users.
Noteworthy this week: the AI fault lines widen
1. AI revenue champions
- Harvey, the AI legal startup, hit $100M ARR in just three years, serving 500+ customers including Comcast, with $800M raised.
- Palantir posted its first billion-dollar quarter, US government revenue up 53%, riding demand for AI-enabled efficiencies.
- Duolingo surged past $1B annual guidance with daily active users up 40% and AI-powered feature expansion beyond language learning.
2. Strategic shifts
- Apple pledged $100B to US manufacturing, expanding AI server and chip production domestically, a geopolitical and supply chain play.
- McKinsey now earns 40% of revenue from AI-related services, deploying 12,000 AI agents to consultants, shifting from pure strategy to outcome-based delivery.
3. Human cost & disruption
- New computer science graduates are struggling to find jobs as AI coding tools replace entry-level developer roles.
- TCS is cutting 12,000 jobs, with other warnings that up to 500,000 outsourcing roles could vanish without reskilling.
In summary: my key takeaway this weekend
GPT-5’s debut shows the next chapter in AI: sophistication, integration, and enterprise deployment. OpenAI’s bet is to make AI the default productivity layer. But capability alone isn’t enough, user experience still wins hearts and adoption.
This week’s wider news makes the contrast sharper. AI is accelerating the rise of companies like Harvey, Palantir, and Duolingo, turning algorithms into revenue and market advantage. But it’s also rewriting the scoreboard in real time, pushing some players off the field entirely.
The lesson? In the AI era, the same force that fuels the winners can just as quickly leave others behind. The future of productivity isn’t just being built, it’s being fought for.
Amitabh Apte 