Building or Integrating with AI?
Let's get you there

Book an advice session to get practical guidance on what to build, how to integrate it, and how to move from idea to implementation.

Book an AI Advice Session

300+ trained · 50+ mentees

Melvin Vivas mentoring a 1-on-1 AI advice call

CLIENT FEEDBACK

“I had a great mentoring session with Melvin. He was incredibly generous with his time and provided valuable insights into the AI industry.”

300+

Trained

50+

Mentees — students, pros & founders

Updated July 4, 2026

Latest AI news for builders and leaders

Fresh artificial intelligence news, model releases, AI agent updates, research, and enterprise AI announcements curated from primary sources.

View all AI news

SecurityJul 2, 2026

Anthropic details Fable 5 cyber safeguards and jailbreak scoring framework

Anthropic published more details on Claude Fable 5 after redeploying the model globally, outlining cyber-safeguard changes and an early draft framework for scoring AI jailbreak severity. The company says it is working with Glasswing partners including Amazon, Microsoft and Google toward an industry-wide standard, and it opened a HackerOne program for security researchers to submit potential Fable 5 cyber jailbreaks for review.

Source: AnthropicRead source

SecurityJul 2, 2026

LawZero publishes Scientist AI safety case for non-agentic predictors

LawZero, the nonprofit AI safety lab led scientifically by Yoshua Bengio, released a formal safety case for "Scientist AI": a disinterested predictor designed to report probabilistic beliefs without pursuing goals of its own. The paper argues that consequence-invariant training can make accuracy and safety reinforce each other, positioning Scientist AI as a possible guardrail for frontier AI systems and as a research accelerator for medicine, climate, cybersecurity and AI safety.

Source: LawZeroRead source

ModelsJul 2, 2026

Mistral releases Leanstral 1.5 for Lean 4 proof engineering

Mistral AI released Leanstral 1.5, a free Apache-2.0 model for formal verification, automated theorem proving and Lean 4 proof engineering. The 119B-parameter mixture-of-experts model activates about 6B parameters per token, is available through Hugging Face weights and a free API endpoint, and Mistral says it saturates miniF2F, solves 587 of 672 PutnamBench problems, reaches 87% on FATE-H, reaches 34% on FATE-X, and found five previously unknown bugs across 57 open-source repositories.

Source: Mistral AIRead source

Meet Your Mentor

25+yrs

300+trained

Book an Advice Call Connect on LinkedIn

Melvin Vivas

Software Engineer & Senior Technology Leader

25 years across banking, fintech, and enterprise.

I've been in the trenches: shipping GenAI apps, building open-source AI tools, and working hands-on with local LLMs and autonomous agents.

I mentor on coding with AI, boosting productivity with AI tools, and turning AI into real business growth. Whether you're a founder, engineer, or leader — you'll get someone who understands both the strategy and the build.

Senior roles at Standard Chartered Bank & Ohmyhome (Nasdaq)

Built AI Dream Photo — production GenAI app

Building aibackends.com — open-source API server

Building skillsbento.com — AI Agent Skills marketplace

Speaker at NTU Singapore, PSIA, and DICT Philippines

Nanyang Technological University, Singapore (NTU) Open Mentoring — Connecting Minds

Invited to speak by these organizations