X to open source all code next month
Next month, every line of code touching the 𝕏 system will be open source and third-party audited.
Saturday, July 25, 2026
X to open source all code next month — Saturday, July 25, 2026
Grok 4.5 solves 30-year graph theory conjecture; X to fully open-source code next month with third-party audit; Apple pilots AI transcription at retail stores. OpenAI launched GPT-Live, a new family of real-time voice models, while Google continues to iterate Gemini with an unlaunched 3.6 Flash variant; meanwhile, new research shows generative AI outperforming humans on select creativity and medical‑diagnostic tasks, and deepfake‑detection systems are advancing beyond facial analysis.
Biggest developments
Grok 4.5 solves open graph theory conjecture
Grok 4.5 just solved a graph theory conjecture that has been open for ~30 years
Samsung bundles Gemini AI Pro with new foldables
Six months of Google AI Pro included with every Samsung Galaxy Z Flip8 and Fold8 device
Apple pilots AI recording at Genius Bar
Apple is piloting a new AI recording and transcription system at Apple retail stores for Genius Bar appointments.
Important posts & threads
Grok 4.6 and 4.7 release cadence
Grok 4.6 in 2 weeks and Grok 4.7 in 4 weeks
Grok 4.5 and Opus 5 lead Pareto frontier
Grok 4.5 and Opus 5 are alone on Pareto frontier
US needs both open and proprietary AI models
i want the US to win in AI both in open source and proprietary models, and i am glad to see this
Bullish takes
- Grok 4.5 benchmark breakthrough accelerates xAI edge
- X open-source mandate boosts platform trust and developer velocity
- Rapid Grok iteration cadence signals sustained model leadership
- OpenAI's GPT-Live launch signals a major step toward real-time, multimodal AI assistants that could reshape consumer and enterprise voice interfaces.
- Google's continued Gemini model iteration (including the registered Gemini 3.6 Flash) suggests aggressive pacing to close the gap with OpenAI and Anthropic in the foundation-model race.
- NVIDIA's dominance in GPU supply for generative AI remains structurally intact as every major model launch and benchmark push still hinges on its hardware.
- Anthropic's focus on safety and reliability continues to attract enterprise and government interest, reinforcing its positioning as a trusted alternative to OpenAI and Google.
- Cloud infrastructure shifts toward AI-optimized regions and GPU-heavy capacity expansions indicate sustained multiyear capex cycles for hyperscalers.
Critical takes
- Regulatory scrutiny of Big Tech's AI and cloud practices is intensifying, with antitrust and data‑governance probes likely to constrain pricing power and integration strategies.
- Google's repeated delays and stopgap Gemini releases (e.g., 3.6 Flash as an unlaunched API string) risk eroding developer trust and enterprise adoption momentum.
- Generative AI's ability to outperform humans on certain creativity and diagnostic tasks raises material workforce‑displacement and liability questions for regulated sectors.
- The proliferation of deepfake‑detection tools like UNITE underscores that synthetic‑media risk is now a core operational and reputational concern for media, finance, and government.
- Venture‑capital appetite for AI startups may cool if benchmark‑driven hype outpaces monetizable use cases, especially in crowded verticals.
Why this matters
Model breakthroughs and transparency moves shift competitive dynamics and regulatory scrutiny in next 30 days.