Minh Pham

[Minh Pham @minhphamdonhat]

@haider1 AI models would eventually settle at how they can get the most out of the intelligence of a model, like how OpenAI and Anthropic have done with Computer Use.

Another focus would likely should keep the model intelligence at around near 5.4 but make it cheaper, more accessible.

[Minh Pham @minhphamdonhat]

@designwkarthick Base on the image and your claim you only use Grok, that means you subagents are using non Grok/Composer models. You have to set default in the settings.

[Minh Pham @minhphamdonhat]

@designwkarthick 60$ cursor already gives you 70$ api credit, 200$ is 400$ api and a pool usage of Grok+Composer; plus, Cursor recently doubled the usage pool twice, that's more than generous! Reason you're burning it because you're using 3rd party models.

[Minh Pham @minhphamdonhat]

@sama Let GPT Voice does the host =)

[Minh Pham @minhphamdonhat]

@walidbelr 3.once Intelligence go beyond common use cases. Ppl will seek cheaper but capable enough to do the work. Priority would shift to reducing costs instead.
=> In conclusion, it's literally same thing that happened to cellular network

[Minh Pham @minhphamdonhat]

@walidbelr I believe the contrary.
1. Rn we are < 50% pple using AI. As more grow, more subsidized => lower price to reach more pple.
2. Intelligence demands will eventually reach its threshold. Just like how we used to have 4K res screen in a tiny phone but nobody cared but go w QHD.

[Minh Pham @minhphamdonhat]

@thsottiaux @thsottiaux Are there pros & cons between the two approaches? If exclude the pricing metrics

[Minh Pham @minhphamdonhat]

@lydiahallie @bcherny @lydiahallie i hope this feedback is useful. There are many use cases we need one agent to owned particular skills or tools, but there's just too much to put into agent definition alone. But also need to ensure the skills won't loaded into other agents which don't need

[Minh Pham @minhphamdonhat]

@lydiahallie I'd like to have extra abstraction to skills and agents. I want to have skills that are exclusive/owned by one/some agents only. Same concept like packaging. Assume skills=function and agent=class and CLAUDE.md is the main().

[Minh Pham @minhphamdonhat]

@lydiahallie This alone let us have blocks that allow us to build reliable and predictable workflows. Let's say... Eh.. Agent Oriented Prompting Language 😂

[Minh Pham @minhphamdonhat]

@lydiahallie Btw Kiro Agent did a very similar concept where it defines agents using JSON and allow imports of skills into agent definition. Would love to see Claude has the same.

[Minh Pham @minhphamdonhat]

@ShadowMonarkV There's a difference. Team Tony most of them already fought Thanos before, so they are prepared of what's coming. Meanwhile, team Cap had no idea what they're dealing with, all they knew about Thanos through words.

[Minh Pham @minhphamdonhat]

@EnzoVerse7 How come nobody thinks they're using the same martial arts 🤔 eh?

[Minh Pham @minhphamdonhat]

@thsottiaux 1. Freedom of choosing whatever harness to use with subscription pool.
2. More range of model pricing and intelligence, from Luna to Sol.
3. Service availability.
4. Reset 🌝

[Minh Pham @minhphamdonhat]

@bcherny This topic is underrated. In the age where everyone literally runs their own agent. Vulnerable models w incapable to resist against PI can potentially causes zero day security issue at large scale.

[Minh Pham @minhphamdonhat]

@bcherny This topic is underrated. In the age where everyone literally runs their own agent. Vulnerable models w incapable to resist against PI can potentially causes zero day security issue at large scale.

[Minh Pham @minhphamdonhat]

@lorxell They were busy dealing with big potato stuffs

[Suryansh Tiwari @Suryanshti777]

Most people want to become AI Engineers

Very few know what to learn next

That's why I'm giving away one of the best Agentic AI Engineer Roadmaps for FREE to the first 4500 people only

Inside you'll learn

✅ Python Fundamentals
✅ LLM Fundamentals
✅ LangChain LangGraph CrewAI AutoGen
✅ LCEL Runnables & Workflows
✅ Memory Systems
✅ Tool Integrations
✅ RAG Systems
✅ Multi Agent Systems
✅ Real World AI Projects
✅ Interview Questions & Answers

72 Hours Only

How to get

Follow me (so I can DM you)
Like + RT
Comment "ROADMAP"

Once the limit is reached I'll stop sending it.

[Minh Pham @minhphamdonhat]

@Suryanshti777 ROADMAP

[Minh Pham @minhphamdonhat]

@vansh22b Same thing with a bagel at 5 bucks, but 3 bagels only cost you 12$, why? Because they either sell it all, or have to throw away everything at the end of the day. The cost of making all bagels are the same.

[Minh Pham @minhphamdonhat]

@vansh22b Let me simplify the math. Assume you have resources that can handle 10M people concurrently. Assume that they don't do reset & 40% users ran out of limits. That's 6M left at the monent in time right? And you know what? 10M and 6M... The maintainence cost are the same. So why not?