Get live statistics and analysis of Ben's profile on X / Twitter

This just changed our Graphic Design pipeline. For context: we build games, have been building our custom AI tools for 2 years now and production is our focus. Latest models by @ChatGPTapp and @GeminiApp are a leap forward. One that we do not experience that often. I created a brand profile for the AlphaGOATs > used it to generate a JSON for this particular piece > generated content. Then different bits can be separated to be updated very quickly with segmenting/inpainting, when needed. Text can be updated using 4o, or just replaced in PS. This is going straight into our AI custom pipelines using OAI APIs. We'll probably have multiple JSON profiles for our different brands, managed by the platform. (Blurring content as some will take this sample for actual details of an upcoming event) Thanks @apollonator3000 for the tip!

Guess who just had a vibe-coding workshop kindly led by none other than @DannyLimanseta ! Engineering, Art, Marketing, Finances, Design, Product all joined! We all had a blast learning from him and we made a juicy Pong clone. Verdict: vibe-coding is real. Thanks, Danny!

Co-founder at Mighty Bear Games, leading AI strategy with an AI-native team. Busy building AlphaAI, the infra behind AlphaGOATs—AI Agents with a persona, that self-promote, compete and earn rewards while you sleep. alphagoats.ai are your 24/7 wealth engine.

Never expected Google to capture the momentum they created with @notebooklm and crank it to 11. Super impressed with the AI native UX they’ve developed and the direction they’re generally taking.

A little late to the party, but it just struck me today how incredibly useful is the @FAL workflow builder to run quick evaluations on prompts, models, etc... We used to do it in ComfyUI. Highly recommend checking it out if you're often updating your image generation stack.

This is the way we operate now: Teams aren’t just building tools, they’re building living, breathing extensions of themselves that plug straight into the company’s Operating System. Same applies to our live Products like AlphaAI, MCP repository for composable ecosystem tools, services, apps... Incredible timing again by @tobi , by now it should be clear to anyone building with AI that MCPs aren’t just connectors, they’re the backbone of a new kind of Operating Systems. In our own experience, here’s what we’re living, not just spotting: Every team has been building MCP-powered tools, Creative Production Agent by the Art Team, the AlphaAI mesh on the Product side, our VibeOps Stack for accelerated internal development, you get the idea. Each tool/interface + associated MCPs are a direct extension of a team’s expertise, packaged as UX/UI and ready to plug into the Company, the Live Product mesh, or sometimes both!! What's been mindblowing to me is that Teams don’t just share data, they share the playbook. Through MCP enabled tools, each team gives access to the instructions, workflows, and know-how, so you work with their data like an insider. It’s like getting an expert in a box. Scout, shared by @tobi is a perfect example of that. LLM flows, knowledge bases, interfaces, databases, they're all compounding through MCPs. Not just linking, but multiplying what’s possible, which is CRAZY. Building this way is now intentional for us, not just accidental. The MCP repository has become the gateway to knowledge and capability, both inside and outside the company. And it's ran by experts in their fields, not replacing anyone. Is this the new standard? Absolutely. I will share more about what we're doing in the near future, there's just too much going on.

[MightyBearLabs]$ log –version 1.5 Let's talk a little about the secret sauce behind Nexus: deep engineering know-how paired with artists prompt engineering. Nexus, our Creative Production Agent, just got even more powerful today! We’ve just shipped intelligent model selection: within a given "tool", Nexus now chooses the right model for the job. First use case lands in editImage tool. Nexus will pick: • Kontext for precision work (crop, zoom, aspect ratios) • NanoBanana for absolutely everything else(for now) Early results are strong even with two models using different schemas, by explicitly documenting parameter deltas in the tool description, Nexus routes cleanly. As we add more, we’ll likely split by tool when schemas diverge to keep the interface sane and predictable. Additionally, and that's the power of an agentic system here, as a user, we can interrupt override the decision by naming our preferred model either in the initial prompt or as we take turns talking to Nexus. The secret sauce: building on top on strong technical capabilities, artists lead the prompt engineering. • @kimtako44 is dialing in guidance prompts for each tool and the chaining strategies between them. • @wemmwem is focused on model specific prompt enhancers that translate intent into production-ready inputs. • Gaute is architecting the agent, driving the technical implementation, and shipping the initial guidance, while our artists iteratively expand capabilities through deep tool use, chaining logic, and workflow extraction. Nexus is already punching above its weight with a small toolset. This is the foundation for a production-grade creative partner that understands craft, not just commands. If you’re curious about where creative tooling is actually heading, not demo-ware, watch this space. We use Nexus daily, as a we build it. What’s next I am not pretending it’s magic: Nexus isn’t yet great at inferring complex workflows from vague, one-line asks. Today, you still get better results if you know what it can do and how to chain requests. As such, it serves the Art team really well. But the goal is for anyone in the team to just use Nexus. So we’re in the process of implementing strategies for rapid context building so the agent can operate with far less hand-holding. The first pillar is Contextual Elicitation, a deliberate tactic to clarify intent fast via targeted questions. Paired with: • smart defaults and assumptions about common creative tasks, • pre-fed known context (projects brand kits, image references, past decisions...), • and learning from prior interactions, …Contextual Elicitation increases Nexus’s success rate and autonomy tenfold. Tighter outputs require much less fewer back-and-forths, leading to faster production. Next up: extending intelligent model selection beyond editImage, standardizing or splitting tools where schemas clash, and evolving chaining guidance so multi-step workflows feel like one ask. We’ll share more soon. Especially about the process artists are using to figure out the changes required to the system prompt, in order to improve the consistency and reliability of Nexus, a task normally reserved for AI Engineers. Ask me anything! And follow @MightyBearLabs for more insights into what we're buiding.

Both @fabianstelzer and @sawyerhood are, in my opinion, casually giving us a glimpse of what the future of UX could be for creative tools. My experience building content with ComfyUI is that the true ROI is when you build very specific solutions around very specific requirements(highly specific workflows). The opposite of generalised solutions. The challenge is creating those solutions at scale when you have 10th of them per project. And, worst, maintaining all these workflows across multiple projects. Well, this is what the experience could potentially look like. And the analogy with Claude's artifacts is spot on: Prompt what you need, preview the workflow, iterate in natural language or by tweaking the workflow directly and repeat until you have the "base" you need. In 10% of the time/effort, it will get you 80% of the result syou need. These are "tools" to make "tools".

Join @MightyBearGames! 🇸🇬 If you are based in Singapore 🇸🇬, we are looking for a Graphic Design Intern to join the team, starting immediately. You have strong Design fundamentals as well as a passion for new tech and visual communication. We are an AI native games team who's worked with Apple, Disney, Razer and are developing our own IPs, infra and AI services. Applications from outside Singapore will be ignored. I'll post the JD link in replies soon, in the meantime, just DM me :)

Massive update from the team at @Scenario_gg ! Mad number of video models available, including Veo3 with ImageToVideo support(yes you read that right). Big big cook!

This just changed our Graphic Design pipeline. For context: we build games, have been building our custom AI tools for 2 years now and production is our focus. Latest models by @ChatGPTapp and @GeminiApp are a leap forward. One that we do not experience that often. I created a brand profile for the AlphaGOATs > used it to generate a JSON for this particular piece > generated content. Then different bits can be separated to be updated very quickly with segmenting/inpainting, when needed. Text can be updated using 4o, or just replaced in PS. This is going straight into our AI custom pipelines using OAI APIs. We'll probably have multiple JSON profiles for our different brands, managed by the platform. (Blurring content as some will take this sample for actual details of an upcoming event) Thanks @apollonator3000 for the tip!

Guess who just had a vibe-coding workshop kindly led by none other than @DannyLimanseta ! Engineering, Art, Marketing, Finances, Design, Product all joined! We all had a blast learning from him and we made a juicy Pong clone. Verdict: vibe-coding is real. Thanks, Danny!

Co-founder at Mighty Bear Games, leading AI strategy with an AI-native team. Busy building AlphaAI, the infra behind AlphaGOATs—AI Agents with a persona, that self-promote, compete and earn rewards while you sleep. alphagoats.ai are your 24/7 wealth engine.

Hey @Taskade , is an agent in Taskade capable of reading the Statuses and other Addons from a given task? Also, can an Agent be working with cross Projects(Workspace level). Currently, I can only create single project agent or teams.

Massive update from the team at @Scenario_gg ! Mad number of video models available, including Veo3 with ImageToVideo support(yes you read that right). Big big cook!

Is the intense praise for @FactoryAI on my timeline paid content? When too many people suddenly praise the same thing, my alarm bells ring. I'm genuinely asking; I have no experience with it.

Rare footage of @Scenario_gg team on a relaxed day at work.

I love @womp3D