Sharing Insights on AI, Online business & Productivity | Helping you leverage AI to grow & monetize | DM for collaboration | [email protected]๐ค
Joined January 2022
- Tweets48.7K
- Following565
- Followers104K
- Likes88K
D-Coder retweeted
A viral Halloween ad used to take a crew and a week.
Now it takes one photo of your product.
Pick one of 93 Halloween presets. Add your product or logo. Done.
Ready before you go to bed.
50% off BuzzHooks, right now.
Introducing the Holiday Buzz Suite.
Viral holiday videos for your product or service, in minutes.
๐ Holiday Director creates viral ads
๐
Mix & Match: pick a scene, character, vibe
๐ 93 Halloween viral presets
More views = more sales. No editing!
Live now, and limited time to get 50% off BuzzHooks for life!
D-Coder retweeted
12 simulation benchmarks.
8 real robot platforms.
One open-source embodied AI harness.
Dockerized environments. Unified HTTP API. Interchangeable models and policies. Plus a 27B preview model.
Technical report: arxiv.org/abs/2610.22471
๐ค Paid partnership
We release the most comprehensive ๐ฒ๐บ๐ฝ๐ถ๐ฟ๐ถ๐ฐ๐ฎ๐น ๐๐๐๐ฑ๐ ๐ผ๐ณ ๐๐ฃ๐ง-๐ฒ-๐๐๐๐ฟ๐ฎโ๐ ๐ฒ๐บ๐ฏ๐ผ๐ฑ๐ถ๐ฒ๐ฑ ๐ถ๐ป๐๐ฒ๐น๐น๐ถ๐ด๐ฒ๐ป๐ฐ๐ฒ across 12 simulation benchmarks and 8 real-robot platforms, together with ๐ฆ๐ฒ๐ป๐๐ฒ๐ก๐ผ๐๐ฎ-๐ฅ๐ผ๐ฏ๐ผ๐ฅ๐ฆ๐: an open-source harness framework and a 27B model that runs at ~1/50 the cost of GPT-6-Astra.
Here's what we found:
๐น ๐ง๐ต๐ฒ ๐ฐ๐ฎ๐ฝ๐ฎ๐ฏ๐ถ๐น๐ถ๐๐ ๐ฐ๐ฒ๐ถ๐น๐ถ๐ป๐ด ๐ถ๐ ๐ฎ๐ฐ๐๐๐ฎ๐น๐น๐ ๐บ๐๐ฐ๐ต ๐ต๐ถ๐ด๐ต๐ฒ๐ฟ. Under a unified protocol (ICL + up to 10ร the interaction budget), GPT-6-Astra achieves clear SOTA across all 12 benchmarks: RoboDojo: 27.6% at 1ร โ 81.9% at 10ร, DexJoCo: 87.3% vs. 40.3% prior SOTA, while the LIBERO benchmarks are pushed to saturation.
๐น ๐ง๐ผ๐ผ ๐บ๐๐ฐ๐ต ๐ฏ๐๐ฐ๐ธ ๐ณ๐ผ๐ฟ ๐๐ต๐ฒ ๐ฏ๐ฎ๐ป๐ด. Even a simple task like placing pens in a holder costs ~$55 in inference.
๐น ๐ง๐ต๐ฒ ๐๐ผ๐ฟ๐น๐ฑ ๐ฑ๐ผ๐ฒ๐๐ปโ๐ ๐๐ฎ๐ถ๐. Time matters as real robots overheat, and an egg burned while the model was still thinking.
Our takeaway: GPT-6-Astra has largely crossed the threshold of ๐ฒ๐บ๐ฏ๐ผ๐ฑ๐ถ๐ฒ๐ฑ ๐ฐ๐ผ๐บ๐ฝ๐ฒ๐๐ฒ๐ป๐ฐ๐ฒ, but ๐ฝ๐ฟ๐ฎ๐ฐ๐๐ถ๐ฐ๐ฎ๐น ๐ฒ๐บ๐ฏ๐ผ๐ฑ๐ถ๐ฒ๐ฑ ๐ถ๐ป๐๐ฒ๐น๐น๐ถ๐ด๐ฒ๐ป๐ฐ๐ฒ is not here yet.
๐ Project page: sensenova.ai/has-gpt-6-astraโฆ
๐ Report: sensenova.ai/gpt-6/assets/doโฆ
๐ป Code & model: github.com/OpenSenseNova/Senโฆ
D-Coder retweeted
Okay, this is where the world-model argument starts to make sense.
The deep learning pioneers from Maluuba are back solving one of the biggest problems in the industry. They unveiled SKY-001, the first engineering world model.
SKY-001 isnโt just generating something that looks 3D.
It simulates different designs, breaks the object down into parts and sub-assemblies, works through the hierarchy, then rebuilds it into something that can actually open in CAD.
Plus, the results compared to frontier LLMs are hard to ignore, including claims of being 150x cheaper and faster.
That feels much closer to how real engineering work happens.
With this release, theyโre also building toward autonomous manufacturing and reimagining what manufacturing firms could look like.
If this approach keeps scaling beyond CAD and BIM, the interesting question wonโt be โwhich LLM is better?โ
Itโll be whether some of these long-horizon problems should be solved with a completely different kind of model.
Introducing SKY-001, the first Engineering World Model.
On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested:
-> 51-125x faster
- > 150-210x cheaper
-> Higher IOU than all frontier models
SKY-001 lets you dream a physical object before you build it.
18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch).
The first โAI General Engineerโ will be the backbone for Autonomous Manufacturing.
D-Coder retweeted
CapCutโs Black Friday promotion is live: 60% off Pro, 40% off Ultra, Oct 20โ30, CapCut Desktop only.
A moody product shot starts the idea; a Halloween ad needs a sequence.
Bring AI-generated storyboards into CapCut PCโs multi-track timeline and develop visuals, text, and audio together.
Review each sceneโs role and refine the flow as the seasonal concept takes shape in motion.
#capcutpc #capcutmarketing #capcutsale #capcutblackfriday
๐ค Paid partnership ยท ๐ค Made with AI
D-Coder retweeted
Your monthly subscription might cost more than your next coffee, but this Dreamina deal is just $0.29 for a full month of Basic. โ
Iโm one of the 1,000 creators officially selected by Dreamina to share this promotion, and I wanted to pass it along.
Remember, the special price is only available through my dedicated creator link.
The next post shows you how to get there.
dreamina.capcut.com/pricing/โฆ
#DreaminaPartner #DreaminaOnly29Cents
๐ค Paid partnership ยท ๐ค Made with AI
D-Coder retweeted
Can stored experience make a coding agent more reliable?
In MemT 1.0-Nanoโs SWE-bench Verified evaluation, average success increased from 66.07% without experience to 69.73% with native memory in the harness-triggered setting.
Tasks solved in all three runs increased from 54.00% to 62.80%.
The result that stands out: more consistent execution when earlier programming experience is available.
huggingface.co/collections/Mโฆ
Can large language models retain information from past interactions and reuse it in the next task?
Introducing MemT-1.0: Towards Memory-Native Evolution.
This time, we've built memory directly into the model architecture.
Our MemT-1.0-Nano and MemT-1.0-Flash make persistent memory part of model computation, carrying information across interactions.
D-Coder retweeted
wait, you can change how it divides the page and keep the rest of the parser?
yep. that's the OXR detail i like: layout analysis is a separate, replaceable model, recognition and reading order stay with OXR, and the 1.3B model still wins olmOCR-Bench, OmniDocBench and DocParsingBench. speed test, 100 pages, lone RTX PRO 6000: 101s vs PaddleOCR-VL-1.6's 144.7s. also the quickest open-source parser they tested, btw.
useful when your application needs a different definition of where one region ends.
github.com/SoMarkAI/OXR
๐ค Paid partnership
D-Coder retweeted
Wait, 122B parameters and it can run on a single NVIDIA DGX?
Thatโs the detail that makes Ornith-1.5-122B much more interesting than another open-weight model drop.
Add MIT licensing, multiple quantized variants, and support across Ollama, Hermes Agent, OpenCode and more, and this starts to feel built for people who actually want to use the model, not just benchmark it.
Aloha!๐บ Meet Ornith-1.5-122B: No Tradeoff Between Strong and Fast
Stronger. Faster. Leaner. Ornith-1.5-122B delivers frontier-level capabilities with exceptional inference efficiency. With superior instruction following, stronger task decomposition, and deeper understanding of real-world workflows, it achieves state-of-the-art performance among open-source models of comparable size, while running on a single NVIDIA DGX card!!
Open weights under the MIT license.
Coding
โ
SWE-bench Verified: 0.845
โ
TB2.1: 0.813
โ
CyberGym: 0.800
โ
NL2Repo: 0.512
Agentic
โ
HLE: 0.446
โ
BrowseComp: 0.785
โ
MCP-Atlas: 0.746
โ
Toolathlon: 0.708
๐คHugging Face: huggingface.co/ornith-ai/Ornโฆ
We canโt wait to see what you build with Ornith-1.5-122B.
D-Coder retweeted
Try it here: tinyurl.com/2xeju4vv
D-Coder retweeted
Imagine you're planning a Tesla Roadster campaign.
You have a visual concept, a mood, and a story you want the audience to experience. How do you turn that brief into a finished marketing video efficiently?
I tested CapCut Web ร Codex as a new way to approach the task.
Instead of treating AI as a standalone generation tool, I used natural-language instructions to communicate the creative goal and let Codex call CapCut to create the video.
This connects creative planning with execution in a more seamless workflow.
For marketers and content creators, that means a more approachable way to experiment with ideas and bring campaigns to life.
#capcutweb #capcutchatgptplugin #capcutcodex #capcutmarketing
๐ค Paid partnership ยท ๐ค Made with AI
D-Coder retweeted
CapCutโs Black Friday deal went live today: get 60% off Pro and 40% off Ultra, on CapCut Desktop only.
The promotion runs October 20โ30.
That trending video where a turning fan blade reveals a different outfit?
You can create it in CapCut and use it in Black Friday marketing. A promising AI-generated shot in the sequence may still have a visual detail worth revisiting.
Local Inpainting in CapCut PC gives me a way to revise that selected area.
I can assess the updated shot within the existing sequence and decide whether it works better alongside the surrounding footage.
Itโs a useful option when a visual needs a focused revision before joining the finished ad.
#capcutpc #capcutmarketing #capcutsale #capcutblackfriday
๐ค Paid partnership ยท ๐ค Made with AI
D-Coder retweeted
Itโs official!
CapCut Black Friday is LIVE TODAY! Get 60% OFF Pro and 40% OFF Ultra, Oct 20โ30, exclusively on CapCut Desktop.
Got a creative product idea inspired by the Jean Phil trend?
CapCut PC helps you turn AI-generated concepts into marketing videos with room for real editing, refinement, and finishing.
#capcutpc #capcutblackfriday
๐ค Paid partnership ยท ๐ค Made with AI
D-Coder retweeted
Viggle's JST-2 can take one full-body photo of you and build the room you were standing in, in 3D, with you in it. then you spin the camera round yourself. going to need a tidier room before I try this
Today we are releasing JST-2, the world's first full-stack physics-native world foundation model.
Many world models learn in pixel/video space. We went back to first principles and built the entire model stack in 3D.
Generate a 3D scene in seconds and expand it in real time. Create characters with motion built in, all through one general architecture.
D-Coder retweeted
Just tried Wizstar for creating Halloween visuals and I'm impressed.
The AIGC templates are super high-quality and let you easily play with different creative looks. Itโs a really fun tool.
Iโll drop a signup link in the comments so you can get 200 credits and try it yourself.
You donโt want to miss the Halloween trend everyoneโs trying ๐
Iโll drop the template in the repliesโand give you free credits to create your own. โฌ๏ธ ๐ป
#wizstarhalloween #halloween #wizstar #aivideo
D-Coder retweeted
if youโre on Ultravox with a different voice provider, nothing changes for you btw. only the built-in Inworld voices got upgraded
Weโre excited to announce that @ultravox_dot_ai is now part of Inworld.
Ultravox is the platform developers use to build real-time voice agents. We've worked with the team for a while through our TTS partnership, and today members of the team that built it are joining Inworld to keep developing it.
