Obsessed with tech. Allergic to untouched ideas
AI, homelab, full-stack. Learning by breaking things
Self-learning as art. Code, languages. The road never ends
GPT Voice has been the undisputed king of speech for years (speech-to-speech index), but lately it’s packing every sentence with “uh”, “well”, “hold on a sec” just to sound fluent, and it honestly feels a bit off.
Then Grok just dropped Voice 3.0 and for the first time it looks like someone is actually challenging that monopoly.
Can’t wait to try it and see if Grok finally breaks OpenAI’s voice lead. 🔥
Introducing Grok Voice 3.0, the world’s most intelligent speech-to-speech model.
It excels at resolving customer calls, responds in under a second, and costs less than other frontier models.
Meta is BACK with a BANG after a whole year! 🔥 Muse Spark 1.3 open weights dropping soon – this is about to shatter China’s year-long open-source monopoly in one move. But the craziest part? They’re not just breaking the wall… they’re igniting the full Agent Robot wars! One mic drop comeback. Who’s ready? 🚀
@TheINAOG The open weights for Muse Spark are coming soon, Muse Wonder models are large models that are too big for many people to use, in addition to potential safety concerns.
Any ai dev, do not miss Vercel’s AI SDK in 2026. I replaced a hand-rolled OpenAI tool loop with ToolLoopAgent. Net diff is basically flat. The bespoke engine is gone. Correctness held. Now we can focus on the core biz without owning the provider client, wire-format messages, JSON-schema projection, or the model/tool while-loop.
Update on the X subscription usage apocalypse 💀
Thanks to @kevin_spacexai for the support tip. Sent a ticket ~12 hours ago. No reply yet, but my 97% usage just got yeeted back to 0% anyway.
X support speedrunning the fix like it’s a ranked match. Efficiency unlocked.
Now testing if this is a real reset or just a temporary truce before the meter starts racing again.
I upgraded from X Premium Legacy to X Subscription Plus this morning. Here are my Day 1 impressions
1️⃣ I heard some users got their usage limits reset after upgrading. Mine didn’t. Is this a bug?
2️⃣ Chat alone seems to burn through my usage WAY faster than Premium Legacy did. And that’s without even touching Build.
3️⃣ I tested OAuth login with both OpenCode and CPA. Both worked, and Build is accessible. But honestly, I’m afraid to use it.
If Chat alone burns through usage this fast, how many prompts will I get with Build before hitting the limit? 💀
Anyone else experiencing the same thing, or is it just me?
Would love to hear your actual usage experience, especially if you’re using OpenCode or CPA with Plus.
For now, I’d hold off on upgrading from Premium Legacy to Plus. At least until we get some clarity on the usage limits.
@tomas_boda Hey Tomas! 👋 Always great to meet fellow builders in the tech community. I’m based in Japan and love connecting with people who enjoy sharing ideas and building cool things. Let’s connect and learn from each other! 🤝🚀
@RaphaelRedmer Hey! 👋 Fellow builder here from Japan 🇯🇵 Always great to connect with people taking the leap to build something of their own. Let’s connect, exchange ideas, and keep building! 🚀🤝
@KI_Vater@X Hey Maurice! 🚀 Always awesome to connect with fellow builders and tech enthusiasts. I believe the best ideas often come from great conversations. Looking forward to exchanging ideas and seeing what we can build! 🤝