What good things is AI doing? Here's 10 things my AI agents found here on X that are examples of good things AI has done for humans. +++++ 1. "ChatGPT saved my life" — grocery store conversation about an ultra-rare diagnosis "Doctors couldn't figure it out. AI did. Turns out
GENERATIVE AI
-
WordPress Categories for AI News Aggregator
By
–
This is my first attempt at subscribers thread. 1/2
-
New Speed and Precision Modes Available in Video Translate
By
–
Try the new "speed" and "precision" modes in Translate:
-
LTX-2.3 Video Model Released with Enhanced Quality Features
By
–
LTX-2.3 by @ltx_model is here
— Replicate (@replicate) 5 mars 2026
→ Crisper fidelity & on-screen text
→ Complex multi-subject prompts that stick
→ Start + end frame I2V with smooth motion
→ Cleaner audio + native portrait + 24/48 fps pic.twitter.com/smohDKlT8TLTX-2.3 by @ltx_model is here → Crisper fidelity & on-screen text
→ Complex multi-subject prompts that stick
→ Start + end frame I2V with smooth motion
→ Cleaner audio + native portrait + 24/48 fps -
Claude Code Solves Internet Connectivity and Browser Performance Issues
By
–
I am on a new wifi and internet provider. Gmail was taking 10+seconds to load, and Jira didn't even load, browser felt slow — even though speed test was returning 500/500 Mbps. I had zero intent to spend hours diagnosing this hell. So I opened Claude Code, described the
-
GPT-5.4 Thinking now available on Replicate with extended context
By
–
GPT-5.4 Thinking is live on Replicate Handles up to 1M tokens of context, best for agentic coding for complex tasks.
-
Grok not as good, Levangielabs way ahead of hyperscalers, says Scobleizer
By
–
Grok isn't even close to as good. I tried. https://
levangielabs.com is way ahead of any of the hyperscalers. Seriously. You don't know, because you don't have access (I'm the first that isn't a big company). -
GPT-5.4 scores 74% on ARC-AGI-2
By
–
GPT-5.4 scored 74.0% on ARC-AGI-2 GPT-5.4 Pro got 83.3%, getting close to a level of Gemini 3 Deep Think.
-
GPT-5.4 Thinking CoT Controllability and Safety Monitoring
By
–
We're publishing a new evaluation suite and research paper on Chain-of-Thought (CoT) Controllability. We find that GPT-5.4 Thinking shows low ability to obscure its reasoning—suggesting CoT monitoring remains a useful safety tool.
-
OpenAI GPT 5.4 Launched: Major Update for Builders and Developers
By
–
Breaking News: @OpenAI GPT 5.4 launched. Just posted a MAJOR update to the report. Now heavily focused on what builders/developers need to know.