Alex Volkov (Thursd/AI) (@altryne) 's Twitter Profile
Alex Volkov (Thursd/AI)

@altryne

✌️ Vibe Coder

🎙️ Host of @thursdai_pod
✨ AI Evangelist with @weights_biases 🪄🐝
working on @weave_wb

Founder and CEO @ targum.video

ID: 5721582

linkhttps://thursdai.news calendar_today02-05-2007 19:19:13

45,45K Tweet

32,32K Followers

1,1K Following

郭明錤 (Ming-Chi Kuo) (@mingchikuo) 's Twitter Profile Photo

My industry research indicates the following regarding the new AI hardware device from Jony Ive's collaboration with OpenAI: 1. Mass production is expected to start in 2027. 2. Assembly and shipping will occur outside China to reduce geopolitical risks, with Vietnam currently the

My industry research indicates the following regarding the new AI hardware device from Jony Ive's collaboration with OpenAI:
1. Mass production is expected to start in 2027.
2. Assembly and shipping will occur outside China to reduce geopolitical risks, with Vietnam currently the
Alex Volkov (Thursd/AI) (@altryne) 's Twitter Profile Photo

I get that Windsurf is part of OpenAI now, but Copilot is part of Microsoft (investor in OpenAI) so why would Anthropic share Opus access with CoPilot but not with Windsurf? 🤔

I get that Windsurf is part of OpenAI now, but Copilot is part of Microsoft (investor in OpenAI) so why would Anthropic share Opus access with CoPilot but not with Windsurf? 🤔
Alex Volkov (Thursd/AI) (@altryne) 's Twitter Profile Photo

Is there a word for UI affordances / buttons that do nothing when user clicks them? Kinda like the "close doors" elevator buttons or... any of these buttons on X

Is there a word for UI affordances / buttons that do nothing when user clicks them? 

Kinda like the "close doors" elevator buttons or... any of these buttons on X
Alex Volkov (Thursd/AI) (@altryne) 's Twitter Profile Photo

For the full show notes and links please see 👉 thursdai.news/may-22 👈 Huge thanks to nisten🇨🇦e/acc Yam Peleg and Ryan Carson for help covering this! We even had a surprise visit from Wolfram Ravenwolf who couldn't resist coming to the Opus watch party! We're also on: 🔗

Alex Volkov (Thursd/AI) (@altryne) 's Twitter Profile Photo

100% with Theo on this one, ganging up on Anthropic and jumping on the "Claude is a rat, anthropic is to blame" is just such an obviously stupid way to disincentivize transparency! All labs have safety testing, all labs should report incidents like this. We shouldn't blame them

Jerrod Lew (@jerrod_lew) 's Twitter Profile Photo

Here is a basic prompting tutorial for Google Veo 3. If you're completely new to the tool, here are some basic ways to start prompting to improve your outputs. We start with basic characters, and build up to multiple dialogues. Check this out.

Alex Volkov (Thursd/AI) (@altryne) 's Twitter Profile Photo

I use GPT 4.5 for writing almost exclusively. 4 Opus has been taking over but not entirely. O3 for every possible web assistant tasks! Reasoning though COT is huge! VEO3 is a drop in for VEO2. I don't get OP point! Saying the big labs should not... continue with progress?

Rob Haisfield (@roberthaisfield) 's Twitter Profile Photo

Alex Volkov (Thursd/AI) Claude 4 actually has a beta header that allows it to reason between tool calls. You can see in the claude interface now multiple reasoning chunks per message sometimes, especially ones that create artifacts.