Vítor Balocco

Your team uses MCP to connect their favorite AI tools to write code, access docs, and query data. MCP wasn’t built for this. We built @runlayer to make MCP enterprise-ready. And raised $11M from @khoslaventures @felicis to enable AI transformation.

What a genius idea. Definitely stealing this

@badlogicgames @zeeg @micLivs Wrong x.com/dsp_/status/20…

@zeeg @micLivs no, the agent generates the ui ad-hoc. in mcp-ui, the mcp server injects an iframe into the agent ui.

Anthropic shipped generative UI for Claude. I reverse-engineered how it works and rebuilt it for PI. Extracted the full design system from a conversation export. Live streaming HTML into native macOS windows via morphdom DOM diffing. Article: michaellivs.com/blog/reverse-e… Repo: github.com/Michaelliv/pi-… Built on @badlogicgames's pi and @DanielGri's Glimpse.

This is so cool and ... it is an MCP App.

MCPs are the opposite of dead. They are the life blood of how AI agents use services inside mid-sized and above companies. Case in point: Uber runs on MCPs internally, for good reason. Details: newsletter.pragmaticengineer.com/p/how-uber-use…

MCP isn’t dead- it’s thriving I’ve seen non-technical teams set up crazy MCP workflows you wont believe It’s not perfect, but it is a great UX

@yenkel Few understand this

if your API has - dynamic client registration via CIMD - user level oauth so agents can auth on their behalf - ability for background agent auth (no UI) - good docs for AI then your API is good enough also, you kind of have an MCP server

They are not dead. They are a must have for any enterprise installation and companies with hundreds of employees. I also thought MCP was dead, but you're thinking like an indie hacker. Assuming you're working at a company with 50 people, people install remote MCP servers, use Oauth and then are done. Everyone is authenticated and you don't have to fiddle around with CLI's. It's also secure. For example check @cursor_ai's new Plugin MarketPlace and you got the idea why MCP's are blooming and why they solve a real issue.

the overindexing on CLIs is kind of insane to me it's building a primitive that's not portable, properly discoverable, has no good approval flow DCR / CIMD to APIs would go so much further but CLI is just the current hype thing

Software engineering was never about the raw lines of code. It’s about managing complexity and fighting the entropy of an ever-evolving system. Many folks painfully (re-)discovering this lately.

@TukiFromKL 2 per day lmao that cute

🚨 Are you paying attention right now? Claude just put a price tag on your senior developer. $15–25 per code review. AI agents that start Instantly The second you open a PR. > Your tech lead makes $200K/year That's ~$400 per code review assuming 2 per day > Claude charges $15 That's a 96% pay cut > And the AI doesn't call in sick on Monday The people protecting their $200K salaries by gatekeeping code reviews are about to have a very bad quarter.

@RhysSullivan Glad you’re finally seeing the light. The main problem with the elicitation mechanism today is how it depends on a stateful MCP connection, which makes it hard to scale — luckily the protocol will get a big overhaul in favor of stateless soon!

I'm hoping that the damage to MCP's reputation gets undone this year Being able to dynamically get input from the user for subtool calls is incredible, and since it's a standard any agent that implements it gets to work with this

@douglasqian @dexhorthy @simonw @boristane

@dexhorthy @simonw Is anyone working on this? Pagerduty for agents? TBH in my experience claude is better than me at debugging system-level failures if you give it the right access to logs, telemetry, etc.

The people I want to hear from right now are the security teams at large companies who have to try and keep systems secure when dozens of teams of engineers of varying levels of experience are constantly shipping new features

If your team is running MCP through @runlayer, your meeting notes are ready.

@boristane Lfg!!

I'm back nominal.dev

@ashtom @cursor_ai @EntireHQ Dying for codex support 🙏 🙏

We’re really excited to add support for @cursor_ai in the @EntireHQ CLI. Experimental for now as the rewind command is not available yet, but other commands (doctor, status etc.) work the same as all other agents. Keep the PRs and feedback coming.

@zeeg @jxnlco it's good feedback sir

I was loving Codex this weekend, and today I've just been fighting the model non-stop. LLMs are still a nightmare of randomness, built on too many conditions that are impossible for you to diagnose. The model and the TUI has improved tremendously though since the last time I used it, and 5.3 is definitely an impressive model. That said, I cant net see a difference in my outcomes between it and Claude Code w/ Opus. I'm sure it performs better at some things and worse at others, but because the technology makes it so difficult to understand whats going to cause the variance, it doesnt feel like a worthwhile optimization to even try to figure out what that is. My biggest gripes are certainly with the UX still. The TUI is a lot faster which is great (albeit I dont like the visual aesthetic), but its missing so many quality of life features that Claude Code has. The MCP experience was pretty good, but everyone messes this up still (dont drop me out of the session just to force me to auth, and do it automatically when needed). I also found myself using plan mode less (to my consequence) because of the difference in UX around it. For whatever reason Claude Code either suggests I use it more (it does), or the habits, but I use it _religiously_ there - and that creates better results. I kept fighting with Codex only to realize I wasn't dropping into plan mode to solve problems. The native /review flow was a nice plus, and I found it more reliable at calling skills entirely implicitly. So a couple nice wins there. Anyways, great model, great progress, its still vim vs emacs tho. Pick the one you achieve the best results with they're all great!

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow. Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes. As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now. It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

@jxnlco @OpenAI @romainhuet @OpenAIDevs What an arc, congrats dude! Huge codex user here, btw

I’ve recently joined @openai to work with @romainhuet on @OpenAIDevs Now is the year of dogged pursuits But Back in 2021 i thought my technical career was over. I had chronic hand pain in both my hands and could barely tie my shoes let alone use my phone or write code. I spent a few years not thinking about what it mean for the value of my labor to go zero but to not being able to produce any labor at all… I gave up bjj. Pottery. Tech. Etc. Then, that one company that solved dota and hide and seek released chatgpt and whisper and all of a sudden with dictation and some determination I could write essays, build things, and make a living from twitter meeting great people like @eugeneyalt @dmdohan @humford @GEVS94 for my reintegration into the tech world after so many years away. From Canada advised companies for free until I had to ask them to pay me. I charged companies until I figured out pricing and asked for enough that I became an investor as well. I started a consulting business and a course business. Learning alongside @HamelHusain and @vig_xyz But through that time I learned a lot about running a business and felt like I’d stopping learning about everything else. I realized that last summer that I wanted to wrap things up and go somewhere and just get involved and be at the center of it all.

@Jordy_vD_ @thsottiaux Not enough cerebras

@thsottiaux I’m on an enterprise plan and would love to use it. Please broaden the rollout

We’ve made GPT-5.3-Codex-Spark about 30% faster. It is now serving at over 1200 tokens per second. More to come on speed across the board.

This week, our team celebrated in Times Square: 5,000,000+ MCP calls secured with @Runlayer. That’s a massive milestone, with a massive billboard. AI is moving at breakneck speed, and we love that. But every new tool introduces new security risks. We’re building Runlayer so no organization has to choose between adopting AI and protecting what they’ve built. 5,000,000+ secured MCP calls means thousands of agents, workflows, and AI products running with real guardrails in place. It means peace of mind for our enterprise customers like Gusto & Opendoor. Huge congrats to the team. And a genuine thank you to our customers.