I thought it was just me! I form my thoughts much more clearly when I type them out than when I try to dictate them to AI. Whenever I use voice mode / speech-to-text it’s like my brain freezes. Only happens talking to AI. Weird.
APPS
-
Character.AI Lorebooks Use Keywords to Trigger Shared World Details
By
–
[explore Lorebooks] Lorebooks let you add world details your Characters can reference – like locations, items, events, species, and more include keyword triggers in your entries when keywords come up in chat, Characters naturally bring that lore into their
-
GPT-Live Handles Travel Tasks While Continuing a Conversation
By
–
A closer look at improved intelligence in GPT-Live: the model can keep a conversation going while helping with multiple tasks at once, like checking flights, pulling up local weather, and shaping an itinerary in real time. pic.twitter.com/hY3wkHVG7z
— OpenAI (@OpenAI) 15 juillet 2026A closer look at improved intelligence in GPT-Live: the model can keep a conversation going while helping with multiple tasks at once, like checking flights, pulling up local weather, and shaping an itinerary in real time.
-
Airtap Aims to Reduce Phone Navigation for Mobile Tasks
By
–
We spend way too much time acting as the middleman between our intentions and our phone screens.
— Charly Wargnier ♨️ (@DataChaz) 15 juillet 2026
Opening an app, navigating five menus, checking out, tracking the delivery etc. you know the drill.@Airtap_ai just launched a new feature that completely bypasses this.
I’ve been… pic.twitter.com/KVtxFthF6EWe spend way too much time acting as the middleman between our intentions and our phone screens. Opening an app, navigating five menus, checking out, tracking the delivery etc. you know the drill. @Airtap_ai just launched a new feature that completely bypasses this. I’ve been
-
A Proposed Interface Separates ChatGPT Chat, Work and Codex
By
–
I don’t disagree at all. I think they should have 3 links in the top dropdown: Chat, Work, and Codex. The chat dropdown should essentially be the old UI and the default when people open the app… that’s what I’d do if I was designing it.
-
Using Projects in Codex While Awaiting Their Return in ChatGPT
By
–
I don’t use GPTs. I do use projects. I just use them in Codex. Saying that, I think they’re hearing the feedback and will get them added back into the app.
-
ChatGPT Work Manages Tasks Across Email, Slack and Calendars
By
–
OpenAI introduces ChatGPT Work, a cloud-based #AIAgent that manages tasks across email, Slack and calendars
by Michael Nuñez @VentureBeat Learn more: https://
bit.ly/4pfHZOg #GenerativeAI #ArtificialIntelligence #MachineLearning -
An AI Shopping Agent Can Request Checkout Inside a Thread
By
–
If a product gates access behind a subscription, a usage top-up, or a one-time unlock, the paywall can live in the thread instead of on a landing page.
— Robert Scoble (@Scobleizer) 15 juillet 2026
A grocery agent builds the cart and sends one request for $40 to check out. https://t.co/pSZkvTnlC5If a product gates access behind a subscription, a usage top-up, or a one-time unlock, the paywall can live in the thread instead of on a landing page. A grocery agent builds the cart and sends one request for $40 to check out.
-
An AI Agent Curates a Personalized X Feed
By
–
I've been building my own personal X algo for the last few weeks, which uses an agent to pre-consume my feed and curate it based on my preferences.
— Gene Kogan (@genekogan) 15 juillet 2026
I just don't think X can deliver the best "For you" experience. The algo is too one size fits all to capture my taste accurately,… pic.twitter.com/JdPyvUZm5dI've been building my own personal X algo for the last few weeks, which uses an agent to pre-consume my feed and curate it based on my preferences. I just don't think X can deliver the best "For you" experience. The algo is too one size fits all to capture my taste accurately,
-
Fable 5 and a Possible Kimi K3 Compare on Universe Simulation
By
–
Kivine, a new model on @arena, is potentially an upcoming Kimi K3 model.
— 🚨 AI News | TestingCatalog (@testingcatalog) 15 juillet 2026
I got lucky and caught a comparison between Fable 5 and Kimi K3 in the Universe simulation prompt.
> Fable 5 finished faster, and most UX components were more robust and easy to use.
> @Kimi_Moonshot K3… https://t.co/5KeH4uNn6T pic.twitter.com/CAP96E6GkKKivine, a new model on @arena
, is potentially an upcoming Kimi K3 model. I got lucky and caught a comparison between Fable 5 and Kimi K3 in the Universe simulation prompt. > Fable 5 finished faster, and most UX components were more robust and easy to use. > @Kimi_Moonshot K3