From product launches to real world impact of AI, we highlight the moments that matter, so you stay ahead || DM for collaborations || 📩 [email protected]

London, England
Joined March 2013
AI is getting out of hand 🤯 AI video tools like Veo 3 have been a huge hit on social media, allowing people to create surreal videos. 10 wild examples: 1. APPLE
662
2,224
815
28,778
17,961,811
Riverside Agents launches Monday: four agents, one per production stage, and nothing posts without the creator's sign-off. Long episodes will be the real test
🤝 Paid partnership
Meet Riverside Agents, the first AI production team built only for video. Producer preps the session, Editor creates the full edit, Clipper cuts the best moments, and Marketer gets it live. They’re experts at video and know how to take your content to the next level. You call the shots. Watch the demo:
5
28
29
3,126
AI Models have come a long way
25
35
7,488
Replying to @pushkersoni72
6) Immersive 3D View: Discover places in a whole new way with 3D models that turn street-level images into cinematic visuals.
1
131
Replying to @pushkersoni72
7) HistoryPin: In partnership with Google, HistoryPin connects the past with the present. Add your historical photos and stories to a global map, and share your legacy with the world.
124
AI Powered 3D Navigation will definitely blow your mind. xitter.cf/pushkersoni72/status/2…
Replying to @pushkersoni72
5) AI Powered 3D Navigation will definitely blow your mind.
1
6
Conversational Search in Google Maps: xitter.cf/pushkersoni72/status/2…
Replying to @pushkersoni72
4) Conversational Search in Google Maps: Users can quickly discover new places and explore detailed information with AI-powered insights.
1
14
Replying to @pushkersoni72
1) Live View Tap the camera icon to start Live View search. Find nearby ATMs, restaurants, parks, and transit stations. See hours and ratings in seconds. Keep everything in one place.
1
8
Replying to @pushkersoni72
2) Lens in Maps Lens in Maps utilizes AI and augmented reality to provide instant insights into your surroundings.
1
13
Context grows. Memory doesn’t. Rock Matrix  open-sourced EdgeRock-4B, a text and image reasoning model with attention memory fixed at 72 MiB from 4K to 256K tokens. Rock Memory Anchors let it recall details long after they leave the local window. EdgeRock-4B achieves 82.9% mean pass@1 on AIME 2024-2026, versus at most 18.8% for the evaluated bounded-state models, including 14B-class models. By combining recurrent state, local attention, and Rock Memory Anchors (RMA), EdgeRock-4B is designed to preserve important information without letting its attention KV cache grow with every token. The result is a different approach to long-context reasoning, built to make memory use more predictable and efficient. For AI developers, this could mean serving more requests on the same hardware while supporting longer reasoning tasks. Feel free to download EdgeRock-4B and give it a try!
7
34
35
4,540
Seedance 2.5 just got real competition on price.  @TopviewAIhq launched Topview T1, its in house video model, now live with Seedance 2.5 level quality for 62% less. You pay $0.22 per second in 1080p instead of $0.585, which puts it at about a third of the cost.  Ultra Annual users also get 30 days of unlimited T1 on top of that. For anyone producing video at volume, this changes the math completely. Side-by-side comparison below:
Introducing Topview T1 🎬 Topview’s newest AI video model. Built to rival Seedance 2.5 1080P quality at 62% lower cost. More shots. More iterations. A much smaller bill. Plus, Ultra Annual users get 30 days of unlimited T1 generation. Put T1 to the test, now on Topview. #TopviewAI #TopviewT1 #AIVideo #AIContentCreation #AIFilmmaking
2
42
120
3,876
The Future: Flying Cybertrucks and Roadsters hovering over a busy intersection.
1
13
54
7,337
Dreamina has never dropped a deal this big: a full month of Basic for just $0.29. Not a trial. Not a typo. We proud to be one of 1,000+ creators officially chosen for it and the $0.29 price only unlocks through an official creator link, so use ours below. dreamina.capcut.com/pricing/… #DreaminaOnly29Cents #Dreamina
🤝 Paid partnership
9
42
107
5,219
Want to recreate a miniature world like this? 1. Find a reference video featuring a miniature scene you love. 2. Upload it to Dreamina Canvas. 3. Use Analyze This Video to break down the prompts and shot structure. 4. Adapt the prompts to your own concept. For example: “A tiny glowing city inside a glass jar, miniature buildings with warm illuminated windows, cinematic reflections on the glass, a miniature train moving through the streets, dreamy mist and floating particles, slow camera push-in.” Use the reference for inspiration, then make the scene your own. #Seedance
1
4
243
You can access the monthly Basic offer through our official creator link: 1. Open my official creator link 2. Scroll down and tap the purchase option. 3. Register for Dreamina or log in. 4. Complete payment at the discounted $0.29 price shown on the official page. One membership month gives you time to explore multiple visual concepts. #DreaminaPartner
4
184
Did you know £49, one time, launches your tool on The AI Library? You get a weekly launch slot with a slot calendar, a countdown, upvotes and leaderboard placement, with eyes on it from founders and investors. Hit Submit your tool in the nav at theailibrary.co and pick your week.
12
61
106
3,381
HERE ARE 38 WEBSITES THAT ARE ACTUALLY USEFUL archive.org — access old content wolframalpha.com — solve complex problems remove.bg — remove image backgrounds tinypng.com — compress images smallpdf.com — edit PDFs ilovepdf.com — merge & split PDFs deepl.com — translate text grammarly.com — improve your writing hemingwayapp.com — simplify writing chatgpt.com — ask questions & get AI help perplexity.ai — AI-powered search notion.so — organize your work trello.com — manage projects canva.com — create designs unsplash.com — find free photos pexels.com — free videos & photos flaticon.com — find icons coolors.co — generate color palettes fonts.google.com — access free fonts namecheap.com — register domains github.com — host & manage code replit.com — code in your browser regex101.com — test regular expressions explainshell.com — understand shell commands fast.com — check internet speed haveibeenpwned.com — check data breaches virustotal.com — scan suspicious files downdetector.com — check service outages 10minutemail.com — temporary email addresses justpaste.it — share text instantly screely.com — create beautiful screenshots carbon.now.sh — turn code into images squoosh.app — compress images similarsites.com — discover similar websites shortcuts.design — find design shortcuts theailibrary.co — discover, try & review AI tools portfoliotab.io — build your portfolio rebrand.ly/startAI — The AI Colony Academy
10
65
129
5,053
51% vs 49%. Would you really call that a clear winner? That's the Hermes Agent vs Grok Bot comparison across 69 posts on @feedbackbench, launching its Personal Agents rankings today. Built by @enterpret_ai, which helps AI teams and their agents identify problems across feedback channels and find proof to fix them, it lets you read real user experiences behind the scores. Check the comparison before picking a side. feedbackbench.com
🤝 Paid partnership
Picking a personal agent? Ask the people who use them. We analyzed 218,149 public posts from 132,844 users across Reddit, X, YouTube and the app stores, and ranked nine personal agents from @AIatMeta, @OpenAI, @GeminiApp, @SpaceX, @NousResearch and others on what users actually say. Meta’s Muse leads overall, with the highest customer love on 22 of 53 criteria. And when people compare Hermes Agent and OpenClaw directly, 83% of those who pick a side choose Hermes. But the best one for you depends on what you need. Want an allowance that lasts? Users rate Muse best. Want to choose your own model? That’s Hermes Agent. Want agents that work as a team? Grok Bot. Rankings update every week. Find yours at feedbackbench.com
1
14
26
3,528
Everyone wants better AI models, but what about fixing the technology running them.  EAGLE-4 is taking a different approach with EAGLE-Harness, which adjusts how tokens are predicted based on the hardware and workload, while its Efficient LM Head cuts draft-model overhead by over 40%.  Instead of just throwing more computing power at the problem, it's finding smarter ways to use what's already there. Curious about the engineering behind it? Dive into EAGLE-4 and see how it works: github.com/eagle-hq/EAGLE-4
EAGLE-4 is here. From EAGLE-1 to EAGLE-3, we’ve kept making inference faster. EAGLE-4 is our fastest yet. 961.6 tokens/s on @Alibaba_Qwen’s Qwen3.8-27B with TokenSpeed (BF16, 4× @NVIDIA B300, TP4, one request, T=0). Github: github.com/eagle-hq/EAGLE-4
3
10
64
19,714