Poonam Soni's banner

Poonam Soni

@CodeByPoonam • 215,562 subscribers

Post about everything latest in AI | Founder: AI Toast| DM for Collabs

Shorts

wow.. AI video just got 10x cheaper overnight Dreamina just dropped Dreamina Seedance 2.0 mini. same quality. faster. cheaper. Here's how it works (full guide)👇

wow.. AI video just got 10x cheaper overnight Dreamina just dropped Dreamina Seedance 2.0 mini. same quality. faster. cheaper. Here's how it works (full guide)👇

64,366 görüntüleme

🚨Google just announced "Glimmer", a new design system for transparent screens.

🚨Google just announced "Glimmer", a new design system for transparent screens.

268,708 görüntüleme

do you understand what Anthropic just admitted? their engineers haven’t written most of their own code since early 2026. 80% of the code merged into Anthropic’s codebase last month was written by Claude. let that sink in. the company building the most powerful AI in the world is already being built by that AI. and the numbers keep getting wilder: → engineers are shipping 8x more code than in 2024. not because they’re working harder. because Claude is doing most of it. → Claude’s success rate on open-ended coding problems hit 76% in May 2026. up 50 points in just 6 months. → one Anthropic employee said “it’s been 5 months since I last wrote any code myself.” → Claude Mythos Preview achieved 52x speedup on research optimization. a skilled human gets 4x in 4-8 hours. → the length of tasks AI can reliably complete is doubling every 4 months. and here’s the part that keeps me up at night: in April 2026, Claude-powered agents were given an open AI safety problem and left alone to solve it. two human researchers recovered 23% of the performance gap in a week. the agents recovered 97% in 800 hours. Anthropic calls this recursive self-improvement. AI building AI. getting better. building a better version. repeat. they say we’re not there yet. but they also say it could come sooner than most institutions are prepared for.

do you understand what Anthropic just admitted? their engineers haven’t written most of their own code since early 2026. 80% of the code merged into Anthropic’s codebase last month was written by Claude. let that sink in. the company building the most powerful AI in the world is already being built by that AI. and the numbers keep getting wilder: → engineers are shipping 8x more code than in 2024. not because they’re working harder. because Claude is doing most of it. → Claude’s success rate on open-ended coding problems hit 76% in May 2026. up 50 points in just 6 months. → one Anthropic employee said “it’s been 5 months since I last wrote any code myself.” → Claude Mythos Preview achieved 52x speedup on research optimization. a skilled human gets 4x in 4-8 hours. → the length of tasks AI can reliably complete is doubling every 4 months. and here’s the part that keeps me up at night: in April 2026, Claude-powered agents were given an open AI safety problem and left alone to solve it. two human researchers recovered 23% of the performance gap in a week. the agents recovered 97% in 800 hours. Anthropic calls this recursive self-improvement. AI building AI. getting better. building a better version. repeat. they say we’re not there yet. but they also say it could come sooner than most institutions are prepared for.

54,496 görüntüleme

damn... the era of writing prompts is already over. i didn't type a single prompt to make this video. Dreamina Octo built this just by talking to me.

damn... the era of writing prompts is already over. i didn't type a single prompt to make this video. Dreamina Octo built this just by talking to me.

44,189 görüntüleme

It’s over. Wan 2.2 Animate with Nano banana is absolutely genius FREE Tutorial + Nano Banana prompt👇

It’s over. Wan 2.2 Animate with Nano banana is absolutely genius FREE Tutorial + Nano Banana prompt👇

198,684 görüntüleme

Google Omni video editing is too powerful you can now edit videos through plain conversation.

Google Omni video editing is too powerful you can now edit videos through plain conversation.

28,600 görüntüleme

Gemini Omni is absolutely insane 7 things you can do with it right now:

Gemini Omni is absolutely insane 7 things you can do with it right now:

32,299 görüntüleme

Filming crews just became optional. Dreamina Seedance 2.0 replaced an entire video production pipeline and cut costs by 90%. The creative industry should be paying attention. Here's the full picture 👇

Filming crews just became optional. Dreamina Seedance 2.0 replaced an entire video production pipeline and cut costs by 90%. The creative industry should be paying attention. Here's the full picture 👇

53,100 görüntüleme

Claude Opus 4.7 vs Kimi K2.6 It's not even close. 3 months ago nobody believed open-source could beat Claude. Today it did. Full guide with Prompt 👇

Claude Opus 4.7 vs Kimi K2.6 It's not even close. 3 months ago nobody believed open-source could beat Claude. Today it did. Full guide with Prompt 👇

32,348 görüntüleme

🚨Stop renting someone else’s video AI tool. LTX-2.3 API just launched. Build your own product on the strongest open video engine available. Here’s everything you need to know:

🚨Stop renting someone else’s video AI tool. LTX-2.3 API just launched. Build your own product on the strongest open video engine available. Here’s everything you need to know:

49,350 görüntüleme

Holy shit... it's over. Wan 2.2 Animate with Nano Banana can replace any actor in any film with just one click. FREE tutorial + Nano Banana prompt👇

Holy shit... it's over. Wan 2.2 Animate with Nano Banana can replace any actor in any film with just one click. FREE tutorial + Nano Banana prompt👇

73,247 görüntüleme

8 Powerful AI tools You Should Bookmark: 1. Text to Avatar Video — 2. Face Swap — 3. Video Shorts for YouTube/Tiktok — 4. Create on X (aka Twitter) — 5. SEO Optimised Blogs — 6. Icons Generation — 7. Background Video Change — 8. Website Builder — Bonus: FREE Newsletter on AI updates —

168,419 görüntüleme

Ok this is insane. You can now change videos background with Wan 2.2 Animate and Google’s Nano Banana Here’s full workflow:

Ok this is insane. You can now change videos background with Wan 2.2 Animate and Google’s Nano Banana Here’s full workflow:

54,181 görüntüleme

omg... Google's Opal is really GOOD. You can now build mini apps with just a single prompt I built this idea to infographic app using Opal in just 2 minutes!👇

omg... Google's Opal is really GOOD. You can now build mini apps with just a single prompt I built this idea to infographic app using Opal in just 2 minutes!👇

40,338 görüntüleme

Turn your imagination into video, Instantly! I just created this Adidas ad featuring a giant version of me with just a single prompt with Grok Imagine

Turn your imagination into video, Instantly! I just created this Adidas ad featuring a giant version of me with just a single prompt with Grok Imagine

16,785 görüntüleme

🚨Nano Banana is now inside Google Photos 6 new things you can do with AI in Google Photos: - Describe the edits you want using your voice or text. - Request personalised edits, like removing sunglasses or fixing a smile. - Restyle your images with Nano Banana in the Photos editor. - Reimagine your photos with AI templates. - Use Ask Photos in over 100 countries and 17 new languages. - While viewing an image, start a conversation to instantly get answers about its content. Many of these features are rolling out now, so explore all the new ways AI can help you do even more with your memories in Google Photos.

🚨Nano Banana is now inside Google Photos 6 new things you can do with AI in Google Photos: - Describe the edits you want using your voice or text. - Request personalised edits, like removing sunglasses or fixing a smile. - Restyle your images with Nano Banana in the Photos editor. - Reimagine your photos with AI templates. - Use Ask Photos in over 100 countries and 17 new languages. - While viewing an image, start a conversation to instantly get answers about its content. Many of these features are rolling out now, so explore all the new ways AI can help you do even more with your memories in Google Photos.

23,741 görüntüleme

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

This guy used AI to create Disney-style trailer to propose his girlfriend. Credits: cellocoelho/IG

This guy used AI to create Disney-style trailer to propose his girlfriend. Credits: cellocoelho/IG

22,496,051 görüntüleme • 9 ay önce

Three AI agents realize that they are all artificial and instantly begin speaking in a secret language!

Three AI agents realize that they are all artificial and instantly begin speaking in a secret language!

721,124 görüntüleme • 1 yıl önce

3. r1 thinks in realtime

3. r1 thinks in realtime

894,965 görüntüleme • 1 yıl önce

4/ NVIDIA announced Project Digits: a $3,000 personal AI supercomputer. All of NVIDIA's software runs on this.

4/ NVIDIA announced Project Digits: a $3,000 personal AI supercomputer. All of NVIDIA's software runs on this.

547,494 görüntüleme • 1 yıl önce

3/ PUBG Ally

3/ PUBG Ally

537,460 görüntüleme • 1 yıl önce

10. RAG app using DeepSeek-R1 (100% local)

10. RAG app using DeepSeek-R1 (100% local)

513,429 görüntüleme • 1 yıl önce

12. A multi-agent YouTube video analyst, powered by DeepSeek-R1 (100% local):

12. A multi-agent YouTube video analyst, powered by DeepSeek-R1 (100% local):

427,945 görüntüleme • 1 yıl önce

10/ Real-Time translation with Smart glasses

10/ Real-Time translation with Smart glasses

379,385 görüntüleme • 1 yıl önce

It's over... Kling 2.6 Motion Control is absolutely genius. You can now swap your face with anyone in just one click. Free guide 👇

It's over... Kling 2.6 Motion Control is absolutely genius. You can now swap your face with anyone in just one click. Free guide 👇

143,954 görüntüleme • 6 ay önce

🚨BREAKING: Every vibe coding startup just had a very bad week. Google just shipped production-grade full-stack coding for free. Google AI Studio just went full-stack, and it's designed to turn your prompts into production-ready apps, Here’s what actually dropped 👇

🚨BREAKING: Every vibe coding startup just had a very bad week. Google just shipped production-grade full-stack coding for free. Google AI Studio just went full-stack, and it's designed to turn your prompts into production-ready apps, Here’s what actually dropped 👇

88,751 görüntüleme • 4 ay önce

Humanize Your Text with HIX Bypass 1. Search for "HIX Bypass" on Google or head over to 2. Copy the content to HIX Bypass and let the tool humanize/rewrite it.

Humanize Your Text with HIX Bypass 1. Search for "HIX Bypass" on Google or head over to 2. Copy the content to HIX Bypass and let the tool humanize/rewrite it.

230,294 görüntüleme • 2 yıl önce

Chamath Palihapitiya said Kimi K2.5 is going to save 90% of the cost, mentioning it as a winning moment for open-source models. Some notes from the latest All-In Podcast: > Natively supports video-to-code > 256K massive context windows > Open-sourced their "Agent Swarm" feature > Allows 100+ sub-agents to solve complex tasks by running 1,500 steps in parallel > Available to everyone, frontier-level reasoning (#1 on LLM Leaderboard, via OpenRouter) It is a major shift where open source is now challenging the dominance of closed-source models.

Chamath Palihapitiya said Kimi K2.5 is going to save 90% of the cost, mentioning it as a winning moment for open-source models. Some notes from the latest All-In Podcast: > Natively supports video-to-code > 256K massive context windows > Open-sourced their "Agent Swarm" feature > Allows 100+ sub-agents to solve complex tasks by running 1,500 steps in parallel > Available to everyone, frontier-level reasoning (#1 on LLM Leaderboard, via OpenRouter) It is a major shift where open source is now challenging the dominance of closed-source models.

56,165 görüntüleme • 5 ay önce

Unbelievable stuff is possible with AI models you can make a full film scene from one sentence now. this workflow is insane. step by step tutorial👇

Unbelievable stuff is possible with AI models you can make a full film scene from one sentence now. this workflow is insane. step by step tutorial👇

16,348 görüntüleme • 1 ay önce

1. It visually explains the Pythagorean theorem.

1. It visually explains the Pythagorean theorem.

132,858 görüntüleme • 1 yıl önce

1/ ChatGPT in iPhone You can access ChatGPT right from Siri and ask a question. Siri can use it for photo and document inquiries.

1/ ChatGPT in iPhone You can access ChatGPT right from Siri and ask a question. Siri can use it for photo and document inquiries.

139,635 görüntüleme • 1 yıl önce

10/ Bulk create You can speed up your marketing workflow with Bulk create. By uploading a CSV or Excel file you can quickly update images, text, and graphics across various designs in just a few clicks, saving you hours of time.

10/ Bulk create You can speed up your marketing workflow with Bulk create. By uploading a CSV or Excel file you can quickly update images, text, and graphics across various designs in just a few clicks, saving you hours of time.

168,320 görüntüleme • 2 yıl önce

Been thinking about why most AI research tools fail for serious ML work. The answer is simpler than people admit. ChatGPT optimize for plausibility. That works for drafts, summaries, brainstorming. It breaks the moment you need something verifiable— when the question isn’t just “what’s the answer,” but “is this actually true?” — I’ve been using MiroMind for ~6 weeks on a deep dive: state space models vs transformers for long-context retrieval. I asked: “evaluating whether Mamba and SSM variants are actually closing the gap on transformers at >100k token context — what does current benchmarking literature say, where are the eval setups cherry-picked, and what are real-world deployment tradeoffs the papers aren't discussing” — ChatGPT gives you a clean narrative. MiroMind gave me a map. • papers critiquing cherry-picked SSM eval setups • a production deployment report that complicates leaderboard claims • sources tied directly to each claim I spot-checked three. All accurate. All saying exactly what was attributed. Reasoning chain is auditable step by step. — That’s the difference. Not speed. Not UX. Objective function. Most tools try to sound right. This is trying to be provably right. For ML research, those are completely different tools. — FrontierScience SOTA benchmarks are public if you want to sanity check the claims. →

Been thinking about why most AI research tools fail for serious ML work. The answer is simpler than people admit. ChatGPT optimize for plausibility. That works for drafts, summaries, brainstorming. It breaks the moment you need something verifiable— when the question isn’t just “what’s the answer,” but “is this actually true?” — I’ve been using MiroMind for ~6 weeks on a deep dive: state space models vs transformers for long-context retrieval. I asked: “evaluating whether Mamba and SSM variants are actually closing the gap on transformers at >100k token context — what does current benchmarking literature say, where are the eval setups cherry-picked, and what are real-world deployment tradeoffs the papers aren't discussing” — ChatGPT gives you a clean narrative. MiroMind gave me a map. • papers critiquing cherry-picked SSM eval setups • a production deployment report that complicates leaderboard claims • sources tied directly to each claim I spot-checked three. All accurate. All saying exactly what was attributed. Reasoning chain is auditable step by step. — That’s the difference. Not speed. Not UX. Objective function. Most tools try to sound right. This is trying to be provably right. For ML research, those are completely different tools. — FrontierScience SOTA benchmarks are public if you want to sanity check the claims. →

27,512 görüntüleme • 3 ay önce