- AI Toast
- Posts
- AI is coming for your Job
AI Won’t Magically Make More Jobs
Image Source: Malak Fouad A| Mo Gawdat
The AI hype machine says work will flourish. Mo Gawdat, ex–Google X chief, says the opposite, and he’s built proof. This is a reality check on what’s actually changing now, not someday.
Mo Gawdat argues that claims that AI creates net new jobs are “100% crap,” citing his startup built by three people plus AI work that once needed hundreds.
Roles across the ladder are vulnerable, from video editors and podcasters to executives; even CEOs face replacement as systems out-execute mediocre leaders.
AGI will outperform humans at most tasks over time; being the best buys time, not immunity.
Employers aren’t only firing; many plan upskilling and internal transitions to help people work alongside AI, delaying mass displacement, not preventing it.
Power users win: learn AI, pair it with sharp soft skills, and compound leverage daily.
A shift in how we view work may follow, with more time for family, creativity, and meaning beyond job titles.
To cushion disruption, Gawdat points to UBI and strong AI governance to curb greed and misuse.
AI isn’t sci‑fi. It’s here, unevenly distributed, and compressing timelines fast. The question isn’t “if,” but “how we adapt” now.
Personal Take: The future favors creative and determined individuals who combine taste, ethics, and AI skills to create unmatched advantages. Build distinctly human strengths, sound judgment, compelling storytelling, and authentic relationships, and then use technology to amplify your impact.
Image Source: Reuters
OpenAI’s GPT-5 launch sparked backlash as users found it “dumber” than GPT-4o. After a Reddit AMA, Sam Altman reinstated GPT-4o, but mostly for paying tiers.
GPT-5 rolled out to 700M+ users, replacing legacy models many workflows relied on.
Users complained about GPT-5’s colder tone and weaker outputs versus GPT-4o.
Altman admitted the new real-time router underperformed, making GPT-5 seem worse. Fixes are rolling out.
OpenAI will show which sub-model answers a query for better transparency.
GPT-4o returns as an option for Plus/Team/Enterprise via “Show legacy models.” Free users miss out.
GPT-4.1 may return only if Plus demand is strong enough.
OpenAI will watch usage to decide how long GPT-4o stays supported.
The main conflict lies between the simplicity of a unified model and the power users' demand for predictable, specialized behavior.
This isn’t just model drama; it’s a product trust moment.
Personal Take: The next wave isn’t a single “smartest” model, but adaptive portfolios with user-governed routing, clear provenance, and SLAs for behavior.
Superhero Tools (Save hours)
Freepik: Free tool to upscale and enhance photos without losing quality.
ChefGPT: AI meal planner that turns your ingredients and goals into quick, personalized recipes.
FitnessAI: Personalized gym workouts that optimize sets, reps, and weights using AI to help you get stronger faster.
Superhuman: Super-fast email client with AI drafting and powerful shortcuts to clear your inbox quickly.
Grammarly: AI writing assistant that fixes grammar, improves clarity, and helps you write in the right tone.
Meet Voyage Context 3
In partnership with MongoDB
If you build AI apps, you’ve felt the pain: chunk your documents to capture details, and you lose the bigger picture. Keep docs whole, and you miss the specifics your users ask for.
Voyage-context-3 by MongoDB fixes this tradeoff.
What it is:
A new contextualized chunk embedding model that reads the entire document at once, then creates an embedding for each chunk that also “knows” the document-level context. So every chunk carries local detail plus global meaning, automatically.
A drop-in replacement. Swap it in for your current embedding model. No downstream workflow changes. Less sensitivity to your chunking strategy, and no need for overlaps, metadata hacks, or LLM-generated summaries.
Why it matters:
Better retrieval, simpler stack: Developers report higher accuracy for both chunk- and document-level retrieval, without extra engineering steps like metadata augmentation or contextual retrieval.
Real savings at scale: With flexible dimensions and quantization (e.g., 2048, 1024, 512, 256; float, int8, binary), you can keep or even improve quality while cutting vector DB storage by up to ~99% in some settings.
Built for long, unstructured content: Legal contracts, research reports, SOPs, financial filings, and technical docs, where the answer often sits in one paragraph, but the meaning depends on the whole file.
Top use cases:
Cross-section Q&A over 50+ page policies
Multi-entity questions in SEC filings
Doc-to-doc comparisons without brittle chunking
Safety-critical lookups in medical/legal workflows
Try it today. Replace your current embedding call with voyage-context-3 and feed it non-overlapping chunks per document. You’ll get a higher recall, fewer hacks, and cleaner code, with context intact.
Quick AI News Bites
Microsoft has rolled out GPT-5 across all its major platforms including Microsoft 365 Copilot, GitHub Copilot, and Azure AI Foundry. The new system automatically picks the best AI model for each task, using faster models for simple questions and deeper reasoning models for complex problems.
India is building AI systems that can work across its 22 official languages and thousands of dialects through government initiatives like Bhashini. The country's conversational AI market is worth $288 million and growing rapidly as businesses deploy multilingual voice bots for banking, healthcare, and retail.
OpenAI CEO Sam Altman has clearly stated that an IPO is not a current priority for the company despite speculation about going public. He criticized the "sketchy" private market trades that investors use to gain exposure to the $500 billion valued company.
xAI announces Grok 4 is now free for all users worldwide in Auto mode, with an option to force Grok 4 via “Expert” and temporarily generous usage limits to explore its full capabilities.
Boost revenue and gain new customers by partnering with us
Reach over 35K AI enthusiasts with your product.
Join our newsletter to connect with tech professionals, investors, engineers, managers, and business owners worldwide. DM now!
Got feedback, a story worth toasting, or a wild tech question? I’d love to hear from you; just reply, or find me on X.
Cheers,
— Poonam Soni