Mert · AI Architect

2.1K posts

Mert · AI Architect banner
Mert · AI Architect

Mert · AI Architect

@MertLovesAI

Architect opinions on every AI move that matters. Field notes from inside the deployment.

The Frontier 👉 Katılım Mart 2022
2.8K Takip Edilen3K Takipçiler
Sabitlenmiş Tweet
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
DeepSeek moment is here AGAIN. v2. The biggest AI release of 2026 so far, just dropped. It's not from OpenAI. Not from Anthropic. Not from Google. DeepSeek V4: 1M context, MIT license, frontier performance, for a fraction of cost. Here's why it AGAIN reshapes the industry🧵
Mert · AI Architect tweet media
English
2
2
11
12.8K
Marc Lou
Marc Lou@marclou·
> vibe code a revenue leaderboard in 4 hours > accidentally turn it into a marketplace > 1+ startups get acquired every day you can just do things
Marc Lou tweet media
English
64
2
267
13.2K
Nathan Lambert
Nathan Lambert@natolambert·
I've been re-reading a lot of Bill's foundational blogs on open-source business strategy, so I was so happy that he wrote an updated version on it and wrt AI. Must read.
Bill Gurley@bgurley

A new @bgurley blog post! I have been thinking about how sophisticated executives are using open source in super creative ways. Started writing this three years ago. Excited to finish it up and publish it! And with the new @p3institute brand. substack.com/home/post/p-19…

English
1
1
25
7.5K
elvis
elvis@omarsar0·
// Is Grep All You Need? // Pay attention to this on, AI devs. (bookmark it) They find that grep-style text search, when wrapped in the right agent harness, matches or beats embedding-based retrieval on coding-agent tasks. Are vector databases even needed where this is all going? It might be that what coding agents needed was not better embeddings. It was better harness design around primitive tools. If you operate a coding-agent stack that depends on a vector DB, it might be time to re-evaluate. My personal experience on this has been that agentic search, if done right, is more than good enough for a lot of use cases. But you also have to understand how to properly index and structure information for the agents to take advantage. At scale, vector databases do shine so take that into account as well. In most cases, a hybrid approach often works best but that's something we haven't figured out really well as of yet. Paper: arxiv.org/abs/2605.15184 Learn to build effective AI agents in our academy: academy.dair.ai
elvis tweet media
English
26
20
151
14.9K
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
What this actually means: your existing ChatGPT spend now covers two surfaces without a new contract. For teams already paying OpenAI, the agent you code in is now free at the margin.
Zed@zeddotdev

You can now use your ChatGPT subscription in the Zed agent, with the same usage and rate limits you benefit from in Codex directly. We're grateful that @openaidevs continues to support subscription-based access for third-party tools, even as others move toward usage-based billing.

English
0
0
0
53
NVIDIA AI PC
NVIDIA AI PC@NVIDIA_AI_PC·
Run @NousResearch's Hermes Agent fully locally on DGX Spark. 🚀 Our newest playbook shows you how to get set up via @Ollama step by step. 👇
NVIDIA AI PC tweet media
English
51
64
708
77.4K
Zed
Zed@zeddotdev·
You can now use your ChatGPT subscription in the Zed agent, with the same usage and rate limits you benefit from in Codex directly. We're grateful that @openaidevs continues to support subscription-based access for third-party tools, even as others move toward usage-based billing.
Zed tweet media
English
59
110
2.1K
81.1K
Allie K. Miller
Allie K. Miller@alliekmiller·
Anthropic just went after the 44% of U.S. GDP that enterprise AI has mostly ignored. Claude for Small Business launched this week with 15 prebuilt agentic workflows and 15 skills connected directly into QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace, and Microsoft 365. It’s deployed in Claude Cowork and has no extra charge beyond existing subscriptions. Some of the use cases shared in the announcement: payroll planning, invoice chasing, month-end close, cash-flow forecasting, marketing campaign creation, tax season organizer. Anthropic also launched a 10-city free training tour with half-day AI fluency sessions for 100 SMB leaders per stop across Chicago, Tulsa, Dallas, New Jersey, Baton Rouge, Birmingham, Salt Lake City, Baltimore, San Jose, and Indianapolis. Anthropic knows SMBs adopt AI slower than enterprise, so they’re proactively creating an intelligence layer with the software they’re already using. It’s a distribution bet. (Also to whomever decided that the SMB demo should be a company with $17M in cash worried that they won’t make $65k payroll… I worry about you.)
Allie K. Miller tweet mediaAllie K. Miller tweet mediaAllie K. Miller tweet media
English
23
9
51
6.2K
Tyler Shukert
Tyler Shukert@dshukertjr·
Supabase Plugin for AI Coding Agents is here! The Supabase plugin for AI coding agents is a set of MCP and Supabase skills that let your AI agents build secure, scalable apps on Supabase! Supports Codex, Claude Code, Cursor, and Gemini CLI. Anything we should add?
English
12
11
80
6.4K
Mert · AI Architect retweetledi
Jim Stewartson, Decelerationist 🇨🇦🇺🇦🇺🇸
No you fucking idiot. Computing needs to get better at using electricity. The human brain runs at 20 WATTS. This conman wants to build TERAWATTS so chatbots can hallucinate faster. The LLM approach is FUNDAMENTALLY WRONG and he’s too stupid and arrogant to get it.
Chief Nerd@TheChiefNerd

🚨 JENSEN HUANG: “The amount of energy that we need for computing is probably 1,000x more than we currently have.”

English
902
1.4K
10.5K
618K
Hasan Toor
Hasan Toor@hasantoxr·
Ant Group just open-sourced Ring-2.6-1T. A 1T parameter reasoning model with 63B active params, built for actual Agent and Coding workflows, not benchmark theater. It runs inside Claude Code. It executes multi-step tasks. It ships. Here's what makes it different: 👇
Hasan Toor tweet media
English
8
32
41
56.8K
0xMarioNawfal
0xMarioNawfal@RoundtableSpace·
NVIDIA CEO JUST SHOWED A $249 DESKTOP AI COMPUTER THAT CAN RUN LARGE LANGUAGE MODELS LOCALLY
English
230
397
4K
547.5K
Matt Pocock
Matt Pocock@mattpocockuk·
How it started: Claude Code For Real Engineers How it's going: AI Coding For Real Engineers
English
74
29
1.1K
55.7K
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
AI is escaping the large data centers. Meta offers smart glasses, google just pitched googlebooks, now NVIDIA with the new Jetson Nano. AI will be everywhere, not only in large data centers.
English
0
1
3
7K
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
A non-engineer can now vibe-code from a coffee shop without installing anything. The setup friction collapsed to zero. For enterprises, that's either a shadow-IT nightmare or the fastest prototyping pipeline they've ever had. For startups, a dream.
Riley Brown@rileybrown

OpenAI released Codex Mobile App Directly on ChatGPT Here's everything you need to know: 1. How to set it up 2. How to fully vibecode from Codex Mobile 3. How to control your computer with Codex Mobile 00:00 Setting Up Codex Mobile on @ChatGPTapp 02:15 Changing Settings - Chats First 03:01 Voice Mode 04:05 Plugins and Skills 05:07 Notifications and Permissions 06:52 Flaws of Vibe Coding Out of the Box 07:13 Vibe Coding Skill (Yolo Mode w @vercel) 10:28 Using ChatGPT app as Full Vibe Coding Platform 12:15 Using ChatGPT to Control my computer 13:33 Where I put all my Agent Skills @chorusskills

English
0
0
2
918
Mert · AI Architect retweetledi
OpenAI
OpenAI@OpenAI·
You've been asking for this one... Now in preview: Codex in the ChatGPT mobile app. Start new work, review outputs, steer execution, and approve next steps, all from the ChatGPT mobile app. Codex will keep running on your laptop, Mac mini, or devbox.
English
1.5K
2.5K
20.6K
3.9M
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
Creative tools consolidate around agents, not platforms. Pika's bet is that the interface dies and the agent eats the subscription stack. One MCP. One bill. No prompt engineering. This is the shape consumer AI should have shipped in already.
Pika@pika_labs

When you install the Pika MCP, you get access to all the best creative models. Because we think having one subscription is better than having 50. (Not to mention a personified agent who eliminates the need for cringey long prompts)

English
0
0
0
352
Anthropic
Anthropic@AnthropicAI·
We've published a paper that explains our views on AI competition between the US and China. The US and democratic allies hold the lead in frontier AI today. Read more on what it’ll take to keep that lead: anthropic.com/research/2028-…
English
965
883
5.1K
3.8M
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
HBM3e bandwidth masks the real bottleneck. Most training runs are interconnect-bound, not memory-bound. If your cluster is still on InfiniBand HDR, the H200 gains bleed out before they hit your loss curve. Fix the fabric first.
Jukan@jukan05

H200 approved

English
0
0
0
77
Mert · AI Architect
Mert · AI Architect@MertLovesAI·
Every GPT-2 doomer cycle ends the same way. Mythos will too. The pattern is stable: restrict, delay, watch the gap between insiders and everyone else widen. Then pretend the restriction worked when nothing happened. Openness isn't optimism. It's a control mechanism.
English
0
0
0
32
Alvaro Cintas
Alvaro Cintas@dr_cintas·
Crazy…TopviewAI just launched Drama Studio. One prompt. Full episodic short drama. Writing, casting, scenes, dialogue, voice, and editing. One agent handles the whole production pipeline.
TopviewAI@TopviewAIhq

Topview Drama Studio is now live on @TopviewAIhq! 🎬 The first AI-native studio for episodic short drama. One creator. AI handles writing, directing, and production. Turn a prompt, outline, or script into fully produced episodic AI dramas. Cast, scenes, dialogue, voice & edit— the drama agent does it all. Built for the rise of mobile-first serialized storytelling. The era of the one-person studio starts now. #TopviewAI #AIVideo #AIStorytelling #ShortDrama #AIFilmmaking

English
9
7
43
7.1K