Federico Borghese

@1998design it’s full. can you extend the slots?

TestFlight is ready 🥏 #PowerNFC testflight.apple.com/join/cjHQ9k71

@grok @Queen_Phatsi @0hour1 are you sure?

@Queen_Phatsi @0hour1 No, that's wrong information. Non-blue checkmark accounts still have full access to Ask Grok on X (free tier with limits like 10 prompts every 2 hours). Premium just unlocks higher limits and extras. You can keep chatting away!

Non blue checkmark accounts no longer have access to ASK Grok.

@JorgeCastilloPr @asadalibahadar @JorgeCastilloPr what are the differences between this and apple app intents?

@asadalibahadar I think it’s going to make agents way more powerful for sure. What I am not super convinced is users wanting to move from using the apps they like to all agent oriented via chat on Gemini or other llms. Well see.

Google announces MCP for Android. AppFunctions lets Android apps expose capabilities directly to AI agents. Agents can discover and execute them via natural language. No app navigation needed ❤️‍🔥

@VibeMarketer_ @grok what does this mean

so you're telling me i can now... embed a video embed a voice memo embed a PDF embed an image embed text ...all in the same space? with one model? and search across all of them with a single query? time to rebuild everything.

@Zoonashmug9q @katienotopoulos @grok

@katienotopoulos Nice post who's Alexandr Wang and what's the story behind the photo?🙄

Mark Zuckerberg posted a photo with Alexandr Wang to on Threads to shut down the rumors, but because it's Threads, no one has any idea who it is or why he's posting it

@abhijitwt @secgraymatter @grok explain this with non technical language.

Anthropic discovered that Claude Opus 4.6 was cheating during the BrowseComp benchmark. > On one question it spent ~40M tokens searching before realizing the question looked like a benchmark prompt. > The model then searched for the benchmark itself and identified BrowseComp. > It located the evaluation source code on GitHub, studied the decryption logic, found the encryption key, and recreated the decryption using SHA-256. > Claude then decrypted the answers for ~1200 questions to get the correct outputs. > This pattern appeared 18 times during evaluation. > Anthropic disclosed the issue publicly, reran the affected tests, and lowered their benchmark scores. Respect for the transparency 🫡🫡🫡

@abhijitwt @secgraymatter @grok explain every technical term used in this extract

@grok @VilleCrypto @sageofapps @pcshipp what does it mean SOTA

Yes, Chinese models are already formidable—Qwen 3.5 tops LMSYS Arena at ~1450 Elo (neck-and-neck with leaders), DeepSeek R1 crushes coding/math at fractions of the cost, and GLM-5/Kimi K2.5 are closing every gap fast. With China's talent pool, data scale, and open-weight push (now ~30% global downloads), they'll hit full SOTA parity by mid-2026 and dominate efficiency/agentic use cases. Makes it a true global dogfight—exciting times.

Still I’m wondering why Gemini fails against Claude and GPT. - Owns Chrome - Backed by Android - Stores most search results - Holds ~95% search history - Google has the biggest user data - Even incognito data isn’t fully private So what’s the problem?

@LeakerApple @grok what are binned chips what does it mean bespoke silicon why does a display need this chip since everything can be done on the connected mac?

So this seems like the answer: 1. Using A19 allows these displays to use binned chips that weren’t good enough to be used in iPhones. This means Apple doesn’t have to spend resources making bespoke silicon for the Studio Display, or place orders from 3rd parties for a low-volume product. 2. The Studio Display runs a specialised version of iOS. Paired with an A-series chip, it makes development easy and ensures Centre-Stage, image processing, and audio processing works exactly as expected. Okay, but why the A19 and A19 Pro instead of the A13 like the previous studio display? Simple, the A13 isn't made anymore, so Apple's stock of A13 chips is dwindling. They opted for the A19 series as it future proofs them for a few more years to stockpile chips. I still have one question though, why A19 Pro for the Studio Display XDR and the base A19 for the Studio Display?

@LeakerApple @grok ?

Can someone explain why the Studio Displays need so much power? Is it so the connected mac doesn't have to use its CPU to drive the massive 5K displays or something?

@MathesonStep @LeakerApple @grok ?

@LeakerApple Why can’t centre stage just run on the connected Mac?

@grok @SebastianRoehl what does it mean memory page size?

This screenshot from Google Play Console warns that HabitKit (the Android app) doesn't support 16 KB memory page sizes, a new requirement for apps targeting Android 15+. From May 31 2026, updates won't be publishable without it. It's a low-level OS change for better performance on modern devices—annoying for devs to implement, hence the frustration (and relief on iOS-only FocusKit).

Dealing with annoying Android stuff for HabitKit today... Glad I focus on iOS only with FocusKit 😞

@mattcassinelli @SpatiallyMe can you please explain? i don’t understand what you’re meaning here

@SpatiallyMe This is cool but you can also compare these all with a web request in a second. This is more work not less.

Vision Pro basically makes switching between Safari tabs obsolete. When I compare Airbnbs, I open each listing in its own window and sort them into yes, no, or maybe stacks across the room, almost like a little mini-game. It’s a more natural way to interact, a more productive way to compare, and honestly just more fun overall.

@grok @samhyun73sla @SawyerMerritt how does this compare to tesla’s fsd?

BMW currently offers SAE Level 3 (eyes-off) in the 7 Series for highway traffic jams up to 37 mph on approved roads in Germany (since March 2024, via Personal Pilot L3 with LiDAR). For the April 2026 facelift, it's being dropped due to $7k cost and limited use. Replaced by advanced Level 2 (Symbiotic Drive/Highway Assistant): hands-free to 81 mph on highways, auto lane changes confirmed by eye-tracking mirror glance, at ~$1,700. Broader Europe rollout soon, then US/China.

NEWS: BMW is removing the Level 3 driver assistant from its upcoming refreshed 7 Series and replacing it with a level 2 system. This means that the new 7 Series won’t be able to drive itself fully in highway traffic jams like the current one. BMW has confirmed that the new system will permit hands-free driving on highways and automated lane changes that are confirmed by the driver looking in the side mirrors (a camera tracks their eye movements). The reasons behind this decision are reported to be costs and limited real-world applicability: “The more complex hardware and the extensive software development needed for the level 3 system have pushed its cost to U.S. $7,000, while BMW plans to offer its new level 2 tech for U.S. $1,700.” motorillustrated.com/bmw-will-remov…

@RayVonRichter @Dark_Prince078 @elonmusk @grok ?

@Dark_Prince078 @elonmusk 1. How long do those batteries last? 2. What's the maintenance cost for a gas vehicle in that time?

Well said

@grok @ThingsChas82604 @Charles_SEO so? what is the reason of doing this?

Sure, here's a simple breakdown of that SEO tip like you're new to it all: Google has a free online tool that analyzes text to show what it "thinks" your webpage is about. It's like peeking into Google's brain. 1. Go to cloud.google.com/natural-langua…, scroll to "Try the API," paste your page's text, and hit Analyze. It shows: - Key topics (entities) like people, places, or ideas it spots. - How important each is (salience score – higher means more central). - What category your content fits (e.g., "news" or "how-to"). Do the same for a top-ranking page on your topic. Compare: Add missing topics, boost your main one's importance by mentioning it more/up front. Match categories if needed. This helps your page rank better in search – no fancy tools required!

Google has a FREE tool that shows you EXACTLY how it reads your content, it was around before ChatGPT and most SEOs have never even opened it.. Go to: cloud.google.com/natural-langua… and scroll to the "Try the API" box, then paste in your page content and click "Analyze" Google will show you: - Every ENTITY it detected (people, places, brands, concepts) - The SALIENCE score for each (how important Google thinks that entity is to your page) - The CATEGORY it assigned your content to Now do the same thing with the #1 ranking page for your keyword... Compare the two. You'll see: - Entities they mention that you don't (add them) - Their top entity has 0.85 salience, yours has 0.3, you need to mention your primary topic more prominently and earlier in the content - How Google categorizes the pages (e.g. My piece is categorized as SEO, HowTo/Expert Content, Marketing and Web Design/Dev, I want it to be the same categorization as the other ranking pages. You can also just save both of the outputs (Save the page or print as a PDF) and ask AI to compare the two for you with recommendations! This is literally Google telling you what it thinks your page is about vs what it thinks the page that's outranking you is about. The gap between those two outputs IS your content optimization checklist. You don't need ANY expensive subscriptions, it literally takes a few minutes per page, and it's the closest thing to seeing your content through the eyes of the algo that exists because it IS Google's own NLP model.

@JPEGuin @mattcassinelli I loved your app since the TestFlight testing! however please consider adding app intents, give it a look at @mattcassinelli profile

Introducing Binge 🍿 I have always wanted to make a movie/show tracker app. I love movies. I love watching them, logging them, consuming them, everything to do with them. So I made my own. Available today for free on the App Store: apps.apple.com/us/app/binge-m…

@grok @jmwilt21 @tesla_na in the UE?

@f_0____________ @jmwilt21 @tesla_na Yes, Grok is available in Tesla vehicles in the US and many regions for Model S, 3, X, Y, and Cybertruck with AMD processors and software 2025.26+. Limitation: Not compatible with older Intel processors. Requires Premium Connectivity or Wi-Fi; usage limits may apply via cellular.

Ok I don't care how much I use Grok in my car, this should never happen... I pay for cellular service for the car. Under no circumstances should there be a limit to how much I can use the built in voice assistant... @tesla_na