Show your work or get left behind.
This episode covers Issue 90 of The OSINT Newsletter and focuses on how OSINT actually gets done in 2026, from investigating developer platforms and deobfuscating messages to building real world credibility without certifications.
In Episode 10 of the OSINT Podcast, host Jake Creps digs into practical investigative techniques and long term skill building. The conversation moves fluidly between tools, methodology, and career advice, showing how investigators can stay effective in a field that changes faster than any syllabus.
Highlights include:
💻 Investigating GitHub Profiles – why contribution graphs can be misleading, how to properly analyze commit history, branches, and activity, and what GitHub can reveal in real investigations.
🤖 Local AI Inside OSINT Tools – a look at tools like God’s Eye that run AI locally using Ollama, enabling private, zero cost analysis without sending data to third parties.
🕵️ Deobfuscating Telegram Messages – how spoiler text and pseudo braille obfuscation works on macOS, and how hidden messages can still be recovered from screenshots and screen shares.
🖼️ Face Recognition and Reverse Image Search – navigating an increasingly paywalled ecosystem and evaluating alternatives for investigators on a budget.
📬 Gmail Addresses as Identifiers – why the ability to change Gmail addresses could weaken email as a long term unique identifier and what that means for attribution.
🍌 Nano Banana vs AI Detection – how new generative models are outpacing detection systems and why AI validation is becoming harder, not easier.
🎓 Why OSINT Certifications Aren’t Worth It – a candid discussion on why certifications often lag behind reality, and what to do instead to build credibility.
Throughout the episode, the emphasis stays on results over credentials, adaptability over static training, and learning by doing rather than memorizing outdated playbooks.
References
God’s Eye | Telegram Spoiler Decoder | Surfface
GitHub Commit History is Misleading | Changing Gmail addresses | Nano Banana vs AI
