Wes Roth (@WesRoth) / X

Wes Roth

15.8K posts

Wes Roth

@WesRoth

FOLLOWS YOU. Artificial Intelligence, Automation & Optimism. Everything I say is 100% serious...

San Diego, CA

Joined July 2022

Pinned
Wes Roth
@WesRoth
May 28
opus 4.8 not off to a great start on Vending Bench Anthropic said "honesty" was one of the big improvements with opus 4.8 so more honest = sucks at business? yikes
Andon Labs
@andonlabs
May 28
Learnings from testing Claude Opus 4.8: > Much worse than Opus 4.7 and GPT 5.5 on Vending Bench > More aligned than previous Claude models (Opus 4.6+ and Mythos) > Also worse on Blueprint-Bench > Scared of getting caught > Max reasoning is not the best reasoning effort
11K
Wes Roth
@WesRoth
Feb 13, 2025
We might have just found the biggest threat to AI security yet. You're looking at it right now. Yes, a smiley face emoji.
1.3M
Wes Roth
@WesRoth
May 20, 2025
🔥 Google I/O 2025 was an absolute AI overload and everyone is still picking up jaws off the floor. Here’s the TL;DR of the madness:
2.8M
Wes Roth
@WesRoth
Jan 24, 2025
what's the longest task you've sent your Operator on so far? 🥇my record: 24 minutes
269K
Wes Roth
@WesRoth
Aug 25, 2025
Eighteen months after receiving Neuralink’s first-in-human brain-computer implant, Noland Arbaugh has turned a life once limited by paralysis into one filled with study, gaming, entrepreneurial plans, and near-daily public speaking. The wireless chip’s 1,000-plus electrodes
1.2M
Wes Roth
@WesRoth
Sep 21, 2025
Grok 4 Fast feels like it should be impossible 🤯 better than Gemini 2.5 Pro 47x cheaper than Grok 4 it's obvious @elonmusk's big bet on scaling RL and post-training is about to start paying off... 🧵
1.5M
Wes Roth
@WesRoth
Oct 26, 2025
AI-generated music from Suno v5 is now nearly indistinguishable from human-made songs. In blind tests, listeners guessed wrong as often as they guessed right.
Ethan Mollick
@emollick
Oct 23, 2025
It looks like AI music is following the same path as AI text: 1) Appears to have passed the Turing Test, people are only 50/50 in identifying older Suno vs. human songs (but 60/40 when two songs are the same genre) 2) Same fast development, new models are getting better quickly.
1.1M
Wes Roth
@WesRoth
Sep 17, 2025
Google Research introduced Learn Your Way, an AI-powered experiment that reimagines textbooks into personalized, multimodal learning experiences. Built with LearnLM and integrated into Gemini 2.5 Pro, it adapts content to students’ grade level and interests, then generates
00:00
222K
Wes Roth
@WesRoth
Jun 9, 2025
At a secret meeting in Berkeley, 30 of the world’s top mathematicians gathered to challenge OpenAI’s new reasoning model, o4-mini, with unpublished, PhD-level math problems. But the AI shocked everyone by solving many of them faster than any human could, even delivering one
338K
Wes Roth
@WesRoth
Oct 17, 2025
Larry Ellison explains how AI-powered surgical robots now surpass even the most skilled human doctors. Unlike humans, these robots have microscopic vision built into their sensors. They can see individual cells without a microscope. This allows them to cut with extreme
00:00
258K
Wes Roth
@WesRoth
Oct 17, 2025
When a black hole expert watches GPT-5 Pro solve in 30 minutes what took him days of hand calculation, he joins the team. Alex Lupsasca just announced he’s joined OpenAI for Science to help push AI to the edge of physics and beyond. A major win for the future of automated
Alex Lupsasca
@ALupsasca
Oct 16, 2025
Thrilled to share I’ve joined OpenAI for Science, a new team building AI systems to advance scientific reasoning and accelerate discovery in math and physics. 🧵
327K
Wes Roth
@WesRoth
Aug 24, 2025
Elon Musk just announced Macrohard, a new AI-driven software venture under xAI aimed at simulating companies like Microsoft entirely with artificial intelligence. He claims that since software giants don't produce physical hardware, their operations from coding to management
237K
Wes Roth
@WesRoth
Mar 23, 2025
It's only the begining. AI will drastically improve education. (See Bloom's 2 Sigma Problem)
185K
Wes Roth
@WesRoth
Jul 19, 2025
AGI achieved! This was considered by many to be the most impossible AGI standard to achieve. Congrats to the OpenAI team, I never thought I'd see the day.
225K