Welcome to the June edition of The Token 😎 This month has been marked by significant developments in the AI landscape, with Claude 4 and Gemini 2.5 showcasing the rapid evolution of AI capabilities.
In this issue, we're excited to introduce our exclusive Innovation Sessions and share insights about the practical impact of emerging standards like MCP in the AI ecosystem 🎯
🤖 The CEO of Shopify seems to have introduced an AI first hiring strategy, essentially opening any position to a potential AI employee before falling back to using humans. Read more.
🤖 Claude 4 arrives with state-of-the-art benchmark scores and integrated thinking modes, reflecting a growing trend in AI systems that can adaptively select appropriate reasoning approaches. Read more
📱 Google updates Gemini 2.5 lineup with new Pro and Flash variants, maintaining their position at the top of chatbot leaderboards. Read more
🎯 Recommendations can benefit from LLMs understanding user preferences better and displaying them in natural language for further feedback from the user. Read more.
📊 Alibaba releases Qwen3 which is their latest flagship model that seems to be behind only o3 and on par with all other frontier models. Read more.
🚀 Mantis introduces an exclusive monthly Innovation Session! In this focused 45-minute consultation, our experts help businesses identify strategic AI implementation opportunities that enhance team performance and drive responsible innovation. Discover how you and your team can optimise the business by using AI.
Interested leaders can secure their spot by emailing bruce@mantis.com - first come, first served. Each session includes a pre-consultation assessment to maximize the value of our time together. Don't miss this opportunity to transform your business with responsible AI implementation!
While not truly transformational, MCP (Model Control Protocol) makes developing AI solutions easier - much like USB-C for connectivity. It doesn't enable fundamentally new capabilities, but reduces the friction of connecting AI models with tools.
The real limitation in AI tool integration has never been about connection standards, but rather about AI models' ability to use the right tool at the right time. And while MCP does solve the M×N integration problem (connecting multiple models to multiple tools), most solutions only need to focus on one carefully chosen AI model connected to specific tools.
Like USB-C, MCP is a useful standard that simplifies development - just don't expect it to revolutionize what's possible with AI implementations 👌
Ready to explore how Mantis can help transform your AI strategy? Book a free consultation with our experts to discuss your specific needs and opportunities: https://mantisnlp.com/contact/#cta 🚀
See you next month! 👋
The Mantis Team
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.