Anna K Clinton : AI (@_kenn50) header
Anna K Clinton : AI (@_kenn50) profile picture

AI Creator • Tech Explorer • Sharing Tools and Insights Building with Brands on daily. 📩DM for Collab [email protected]

Miami, FL
Joined October 2022
Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)Photo by Anna K Clinton : AI (@_kenn50)
NO WAY OUT - One Man, One Building. GPT Image 2 x Seedance 2.5 | @dreamina_ai He walked in alone. Two floors, a small army, and a boss in a white suit standing between him and the top. One man against all of them. Made with Dreamina Seedance 2.5.
2
29
1
57
23,178
The next big leap for AI agents might be solving engineering problems, not just generating text, images, or code. Archimedes explores and improves physical designs, while Reynolds runs numerical physics simulations to evaluate their performance. The examples are pretty interesting: A redesigned datacenter layout targeting $272K in savings. A lid-free mug that spilled 42% less in simulation. A turbine blade with 25.5% lower aerodynamic losses in a 2D benchmark. Reynolds (Open Source, MIT Licensed) can also turn plain-English prompts into simulations with inspectable results. AI isn't just helping create designs anymore. It's helping engineers determine which ones perform better.
We are Introducing an AI engineer that turns physics simulations into better designs It redesigned a data centre’s cooling system to use less electricity and found a turbine blade that beat a leading published design Today we’re open-sourcing the simulation agent behind it
24
11
121
55,576
we spend weeks right sizing infra and then send every LLM call to the biggest model. RouterPlus kinda exposes how dumb that is
AI teams overpay for inference twice: once for models bigger than they need, and again for idle capacity they don’t use. @useRouterPlus is the extended inference team for fast-growing AI companies. We test models and providers on your traffic and get providers to compete for your work. Make your inference better. Go to routerplus.com
14
17
70,354
Tesla Roadster on a winding mountain road. Golden-hour lighting. A cinematic opening shot.Sometimes a video starts with just one scene in your head. The CapCut plugin is coming soon to ChatGPT, giving you a conversational way to explore that idea with CapCut Web.Early Access is currently available through ChatGPT desktop (PC only). #CapCutDidThat #CapCutWeb #ChatGPT #CapCutWorld
1
13
20
55,004
Anna K Clinton : AI retweeted
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
930
714
244
11,093
5,223,558
GPU: Please stop asking me to generate another 10,000 tokens. 😭 EAGLE-4: What if we made decoding smarter instead? Same hardware. More potential. This is the kind of AI infrastructure progress that deserves more hype. 🚀🔥
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
1
45
137
28,723
The AI industry has been chasing bigger models for years, but inference efficiency is becoming just as important as raw intelligence. Delivering competitive performance with more manageable hardware requirements is a meaningful direction.
Aloha!🌺 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 ✅SWE-bench Verified: 0.845 ✅TB2.1: 0.813 ✅CyberGym: 0.800 ✅NL2Repo: 0.512 Agentic ✅HLE: 0.446 ✅BrowseComp: 0.785 ✅MCP-Atlas: 0.746 ✅Toolathlon: 0.708 🤗Hugging Face: huggingface.co/ornith-ai/Orn… We can’t wait to see what you build with Ornith-1.5-122B.
1
20
103
49,175
"my speech at my sister's 1. "my speech at my sister's wedding everyone loved it. I hate rewatching it because I look like I'm about to cry the whole time smoothed out my own face with @trydrama_io. kept the speech. kept the tears at the end, those were real"
Introducing our new AI model @trydrama_io to kill reshoots and make acting editable! Control what part of a performance to preserve and what part to change using a simple intensity slider. We just nailed 3 industries(thread):
3
6,729
Anna K Clinton : AI retweeted
CTRL+F is not enough anymore. Imagine asking a question across hundreds of PDFs, financial reports, spreadsheets, and contracts. And getting an answer backed by the original documents. That's the idea behind combining three AI layers: NDI → Understands the documents. Drex → Makes decisions using extracted information. GPT → Reasons over the results and explains them. NDI supports 36 file types and is tied for #1 on the parsing leaderboard. Search is evolving from finding keywords to understanding information.
Introducing NDI 1.0 - a small model for document processing. #1 on Parse Index Normalized (across 6 public benchmarks) GPT-6 Astra performance at 1/10th the cost 📈 Trained on 15M+ financial files for industry-grade accuracy. Get started with $25 in free credits: nace.ai/ndi Reply "NDI" for a single file setup for Claude Code, Codex, PI, Opencode, Hermes, Prime Agent, and more. Start parsing and classifying documents.
24
9
1
129
111,137
The real breakthrough is reducing the gap between having information and making informed decisions. Businesses already have enormous amounts of data; extracting useful intelligence from it is the difficult part.
CTRL+F is not enough anymore. Imagine asking a question across hundreds of PDFs, financial reports, spreadsheets, and contracts. And getting an answer backed by the original documents. That's the idea behind combining three AI layers: NDI → Understands the documents. Drex → Makes decisions using extracted information. GPT → Reasons over the results and explains them. NDI supports 36 file types and is tied for #1 on the parsing leaderboard. Search is evolving from finding keywords to understanding information.
1
2
4
115
Love how restrained this is. It doesn’t need flashy visuals or constant cuts to hold your attention — the emotion does all the work.
PSA: Time to call someone. My new short, "Hello?", written with Sherpa by Pocket FM.
1
1
4
232
Anna K Clinton : AI retweeted
What makes an AI investment consultant feel trustworthy enough for high-net-worth clients? 🤔 I’ve been testing Wizstar’s interactive AI avatars, and it got me thinking about how AI could be used in private fund investment consulting. For high-net-worth audiences, professional credibility is everything. It’s not just about providing information, but about delivering it in a clear, consistent, and professional way. With Wizstar, an AI avatar could act as a virtual investment consultant, helping introduce private fund strategies, explain investment concepts, answer common questions, and guide clients through different scenarios in a more interactive way. Imagine having a professional digital consultant available 24/7 to support client communication, initial consultations, product introductions, and investor education. 👀 I’m testing the experience to see how far AI avatars can actually go in real-world financial scenarios. Here’s what I tried: wizstar.com/ai-avatar #AI #AIAvatar #DigitalHuman #Investment #FinTech #PrivateFund #Wizstar
🤝 Paid partnership
247
59
3
122
58,216
Wrong Corner, Wrong Guy Three against one, a dark corner, and no way past them. So he rolls up his sleeves, takes a breath, and shows them exactly why they should've picked someone else. Made with @dreamina_ai Seedance 2.5.
🤖 Made with AI
240
44
56
13,495
PROMPT: man = a young man around 25, fit and composed, casual sharp streetwear — a fitted jacket, jeans, sneakers — short neat hair. Calm, confident, quietly capable. Appearance only. Strangers = three rough-looking men in dark casual and hooded clothing, menacing, moving as a threatening group, empty-handed. Appearance only — a template repeated with variation. Street = a quiet city street at night, rendered photorealistically — dim wet pavement, glowing streetlights and shop signs, a dark corner where a side street meets the sidewalk, moody shadows, light mist, deep urban night atmosphere. Cool desaturated blue-grey palette with warm streetlight pools. Environment only. SCENE: a tense cinematic night scene. A young man walks alone down a quiet city street at night; as he rounds a dark corner he is suddenly confronted by three men who move to threaten him. Unshaken, he calmly folds up his sleeves, takes a steadying breath, and then defends himself against all three at once — fast, skilled, and controlled — taking them down. Cool, moody, intense. TECHNICAL: 16:9, cinematic night action, cool desaturated blue-grey grade with warm streetlight glow, dynamic handheld camera, moody low light, wet reflections, motion blur, fine film grain, photorealistic gritty quality. CUTS: CUT 1 (0-6s): The young man walks alone down the dim night street, hands in pockets, relaxed, streetlights pooling on the wet pavement. Smooth tracking shot behind him as he approaches a dark corner. CUT 2 (6-11s): He rounds the corner — and three menacing men step out and block his path. He stops. A tense standoff, the men sizing him up. Low angle building the threat. CUT 3 (11-16s): Unshaken, he calmly folds up one sleeve, then the other, rolls his shoulders, and takes a slow steadying breath — a quiet, confident moment before the storm. The three men hesitate. CUT 4 (16-25s): The confrontation erupts — he moves into fast skilled self-defense, taking on all three at once: dodging, striking, blocking, using precise controlled movements to drop them one by one across the wet street. Dynamic changing angles, intense and fluid. CUT 5 (25-30s): The last man drops. The young man stands alone amid the three, straightens his jacket, exhales, and calmly walks on into the night. Cool confident final frame. END. RULES: References are appearance only, do not recreate. Keep the young man's appearance consistent across every cut. Action is intense but not gory — no blood or graphic wounds, impacts sold through motion and body mechanics. Cool moody night grade throughout with warm streetlight pools. Sound design: quiet night ambience, distant city hum, the tense standoff, impact hits and grunts during the confrontation, a driving tense score, ending on his calm exhale and footsteps. Final frame cool, stable, and clean.
1
1
2
454
This feels like a strong example of where conversational avatars could provide real utility. Not just presenting information, but creating an interactive layer between clients and financial institutions.
What makes an AI investment consultant feel trustworthy enough for high-net-worth clients? 🤔 I’ve been testing Wizstar’s interactive AI avatars, and it got me thinking about how AI could be used in private fund investment consulting. For high-net-worth audiences, professional credibility is everything. It’s not just about providing information, but about delivering it in a clear, consistent, and professional way. With Wizstar, an AI avatar could act as a virtual investment consultant, helping introduce private fund strategies, explain investment concepts, answer common questions, and guide clients through different scenarios in a more interactive way. Imagine having a professional digital consultant available 24/7 to support client communication, initial consultations, product introductions, and investor education. 👀 I’m testing the experience to see how far AI avatars can actually go in real-world financial scenarios. Here’s what I tried: wizstar.com/ai-avatar #AI #AIAvatar #DigitalHuman #Investment #FinTech #PrivateFund #Wizstar
1
135
Going beyond ad classification into customer journey stages, creative styles, and full account analysis makes this much more useful than a basic ad-library scraper.
JEV is insane for competitor research! We gave it Resilia's ad library It classified 1,891 ads in 19 seconds for $0.12 costumer journey step + Ad style And a complete deep analysis of the account Coming soon to maxfusion MCP
3
11,141
Anna K Clinton : AI retweeted
I’ve been testing different AI tools for e-commerce tasks, including Accio and comparable tools like GPT/Codex and Claude. What stood out to me wasn’t just performance — it was how much that performance costs. Across Accio’s 107-task e-commerce benchmark: Accio: ~$3.69 OpenAI Codex: ~$9.27 Claude Code: ~$9.51 From my testing, all of these tools can be capable, but Accio’s approach to cost-efficiency makes a lot of sense for e-commerce. Instead of using maximum compute for everything, it matches resources to the task — keeping routine work lightweight while using deeper reasoning where it’s actually needed. For me, that’s the bigger takeaway: high efficiency doesn't have to mean high AI costs. Accio becomes the most cost-effective E-commerce AI tool for SMEs.
🤝 Paid partnership
29
24
67
55,710
Anna K Clinton : AI retweeted
What caught my attention isn’t just GPT-6 Astra topping the TRACES leaderboard, it’s how it was evaluated. TRACES measures discovery across 6 capabilities: Tools, Repair, Alternatives, Coherence, Evidence & Scope. Astra made major gains in some while others stayed steady. That’s why Discoverative AI is interesting: it’s not just about what AI knows, but how well it can discover what it doesn’t.
🤝 Paid partnership
GPT-6 Astra @OpenAI is here. Everyone's asking how smart it is. We asked a different question: can it discover the unknown? We ran it through TRACES — our benchmark for scientific discovery capability. Six process dimensions, not one score. Result: it tops our leaderboard. Real generational jump — Evidence and Coherence both up ~20-25% over GPT-5.6 Sol and GPT-5.5. Highest Coherence we've ever recorded. But the six dimensions don't move together, and that's exactly why one "how smart" number isn't enough. Some capabilities leap forward, others hold steady, in the same model. That unevenness is what a process score is built to catch. Full breakdown, live: discovertraces.ai/liveleader… #GPT6 #TRACES #AIforScience #LLM #AIBenchmark #Apodex #AGI #AIResearch #ScientificDiscovery #FoundationModels
3
20
28
11,937
Anna K Clinton : AI retweeted
Turning a movie into a handcrafted felt world sounds like a lot of work. Fortunately, the **Flova Agent** likes doing the work. @flovaai Give it the IP, the style, and a few references—it helps plan and create the whole visual story. Prompt shared below.
🤝 Paid partnership
39
17
71
17,697