Lena (@_LenaHart) header
Lena (@_LenaHart) profile picture

Trying AI tools in real workflows and sharing what actually helps. Notes on creating, building, and working a little smarter.

Joined February 2013
Lena retweeted
The gap is what caught my attention here. The best AI scores 0.76 correlation on Tumor Context. Repeated lab experiments reach 0.97. Promising progress, with a measurable gap still to close. This is the kind of grounding I want to see in AI for biology.
🤝 Paid partnership
Cancer is not one cell. It is a neighborhood. Together with @InventiaLifeSci, we are launching Tumor Context, a new benchmark in BioSimBench: can AI learn the grammar of the tumor microenvironment? celltype.com/biosimbench/tum…
6
66
62
20,904
Lena retweeted
Precision is no longer manual. From placing the first CPU to running final AI vision inspections, automated engineering is setting a completely new standard for hardware reliability.
1
1
240
Change a tumor’s neighbors. Can AI predict what happens to its genes? That’s the test behind CellType × Inventia’s Tumor Context benchmark, using real 3D lab experiments. I like this direction: testing whether AI can predict biology when the context changes
🤝 Paid partnership
Cancer is not one cell. It is a neighborhood. Together with @InventiaLifeSci, we are launching Tumor Context, a new benchmark in BioSimBench: can AI learn the grammar of the tumor microenvironment? celltype.com/biosimbench/tum…
8
53
49
14,724
My bar for any “AI team” product: can another human join the work, or is it still just one person managing a collection of bots? Teamily AI is built as a Human–Agent Platform, so I tested it with a partner and a decision that needed receipts: Choose my next tutorial, verify that viewers could try it for free, then turn the recommendation into a usable comparison page. The research agent compared three possible tutorials: fact-checking a post, creating a voice summary, and making a slide deck. Its first recommendation was the fact-checking tutorial. I challenged the missing sources and access requirements. It went back through the evidence and changed its recommendation to the voice summary. Then Webpage Developer used that revised research to build the comparison page with Website Builder. I commented directly on the page: add a free practice exercise and make the tool limits clear. My partner added another requirement: a “Before you start” checklist so beginners would know what they needed. The developer incorporated both changes into the finished page. That was the real test for me: two people and multiple agents challenging, handing off, and improving the same piece of work from decision to delivery. I still made the calls. The group carried them through. Give @Teamily_AI a decision worth challenging 👇 teamily.ai/i/3PG6Q7 Open the finished comparison page: teamily.ai/link/nDWcfeVrjPM
150,000 small business owners and employees are already using Teamily AI. Today, we're opening it to everyone: your personal AI agent and an entire team of AI agents, ready to work for you and your business. Meet Teamily AI 2.0 — the human–agent platform where you and always-on AI agents get work done together in one unified, context-aware space. We've taken four major capabilities to the next level to help businesses build and grow faster: 1. Real-Time Human–Agent Collaboration. Humans and AI agents work together across chats, meetings, studios, and communities to research, analyze, and create websites, dashboards, videos, slides, documents, and more. 2. Unified Context & Living Memory. One connected context across conversations, meetings, channels, and workspaces. Your AI agents remember, learn, and understand you over time—delivering smarter, more personalized results. 3. Custom Agents & Agent Teams. Create and orchestrate AI agent teams in one click. Manage agents through Agent Workspace, connect them to Telegram, WhatsApp, iMessage, and Slack, or integrate them into your products via Agent API. 4. Intelligent Multi-Model Routing. Our proprietary models work alongside leading models from OpenAI, Anthropic, Gemini, DeepSeek, Qwen, Kimi, MiniMax, and more—automatically routing each task to the best model across text, image, video, and voice. Teamily AI 2.0 is built for every business owner, every employee, and every customer. America has 36 million small businesses. 98% have fewer than 20 employees. Many are teams of one. At Teamily, we believe AI should democratize opportunity, giving everyone the power to achieve more. You have the business. Now you have the team. Try Teamily AI 2.0 for free → teamily.ai
29
121
128
63,679
“Warm, textured walls” could mean sandy plaster to me and rough stone to someone else. I’d use Maket to try both on the wall and see which one actually fits the room.
🤝 Paid partnership
Generate floor and wall textures from text or an image. Light oak, terracotta tile, or checkerboard marble? Try it out in your space and see what feels right. Your next room refresh starts in Maket.
17
85
86
18,313
One Meta ad spent 26% more and got fewer clicks. Would I cut its budget? I used Moby to see what the account data could tell me before making that call. I tested it in @triplewhale’s demo account using GPT-6.1 Sol. First, it compared two weeks of Meta campaign performance. Then I asked it to investigate the rising click costs at ad level. It identified an ad where spend went from $2,817 to $3,554 while clicks fell from 3,234 to 3,131, and laid out what to inspect next. What I liked: I could keep questioning the findings against the connected data. Higher CPC alone wasn’t enough to justify a budget change. triplewhale.com/?i=kaiteeshi… #triplewhalepartner
🤝 Paid partnership
15
90
89
42,139
CapCut Pro creators, the Dreamina × CapCut collab just dropped something massive for your next video project. 🎬 You can score up to 10 free Seedance 2.0 generations, extra bonus credits, and snag a month of Basic for just $0.29 if you are a new Dreamina user. Every single eligible user walks away with rewards. 🎁 dreamina.capcut.com/ai-tool/… #DreaminaPartner
16
92
93
41,869
Five people gave feedback in three places and now it's your job to merge it. The producer wrote in Slack. The client answered the review email. Your teammate left a note somewhere you haven't found yet, and two of them disagree about the same shot. YouArt Desktop puts all of it on one canvas. Notes from Slack and Gmail land as comments pinned to the shot they're about, and teammates and clients can comment right there on the work too. The agent reads those notes and clears them one by one, next to you. You get to spend the meeting looking at the actual video.
🤝 Paid partnership
13
90
93
38,392
A bug mentioned on a customer call can sit there until someone turns it into a task. That’s the part of @vorflux 2.0 that caught my attention: AutoBots pick up the work, track it and bring in the right person for approval. I’d try the Customer Experience bot first.
🤝 Paid partnership
Today we're launching @vorflux 2.0: AutoBots, 24x7 coworkers that live in Slack. I previously co-founded @Rippling, now $10B+, engg team of 1000s. Engg changes crazy fast now. IDE autocomplete > CLI > Cloud agents > multi-player slack agents Models deliver predictably now. Bottleneck to parallelism is to initiate, keep track and ping the right people for approvals. Hire from our directory of opinionated AutoBots or make a new one with a goal. Bring your Codex/Claude Code subscriptions and Vorflux is practically free. We recently closed a $15M seed round for @vorflux backed by @ycombinator, @peakxvpartners, @alliancedao, along with incredible partners including @parkerconrad, @immad, @typesfast, @jake_zeller, @balajis, @nivi, @metakovan, @lmrankhan, @nikitabase, @0xrwu, @ayushjaiswal, @mattshumer_, @eshamanideep, @sreeramkannan, @dvcoolster, @nusimow, @TeddySolomon11, @ashtoncofer, @rvivek, and others. Lmk which autobot you’d like to try and we’ll double your credits to try. Chapters: 0:00 Introducing Vorflux 2.0 0:22 From manager to director 0:44 How an AutoBot works 1:04 Cloud setup 1:25 On-call AutoBot 1:42 Project manager AutoBot 2:03 Customer experience AutoBot 3:01 Designer AutoBot 3:52 Social AutoBot 4:51 Open and frontier models 5:23 Always-on agents
7
49
56
21,465
I have product photos for everything in my store. Going to see how much B-roll I can get out of them this weekend.
🤝 Paid partnership
Product shots in. A full set of B-roll out. Generated with Fable 5.2 and the ElevenLabs MCP.
1
47
51
21,006
A wave hitting a coastal town, in an IFM demo... did not expect to hold my breath
🤝 Paid partnership
Introducing PAN 2: an open foundation world model for physical intelligence. A language model predicts the next word. PAN 2 predicts the next world: what happens after each action, grounded in time, physics and space. Towards AI that simulates “what if?” before it acts. More details: ifm.ai/pan/
60
66
17,741
Honestly my favorite part of The Search was scrolling the entries. So many jingles saved. Congrats to the winners!
🤝 Paid partnership
First place in The Search goes to “Sambal!” by Brillian Fairiandi, taking home $50,000. Congratulations! Thank you to everyone who entered, and a big congrats to all eleven who placed. See the full list here: elevenlabs.io/creative/these…
54
58
15,851
My team already dumps files, pics and half-baked ideas into a group chat all day, so having Teamily AI's team agents join that same chat and work off all of it makes a lot of sense to me
🤝 Paid partnership
Messaging Connected People. AI-Native Messaging Will Connect the Agentic Internet. The first generation of instant messaging connected people. It changed how billions of us communicate, socialize, and work together. But the next generation of the Internet won't be built for humans alone. AI agents are becoming our teammates, creators, and service providers. To truly work alongside humans, they must communicate and collaborate as naturally as people do, while going far beyond human capabilities with richer context, living memory, continuously self-improving intelligence, and seamless coordination across people, agents, and services. Yet today's messengers were designed for human-to-human communication, not human–agent collaboration. They lack the native capabilities to support intelligent, context-aware human–agent teams. It's time to reinvent messaging for the Agentic Internet. That's why we're reinventing instant messaging at Teamily AI. 👥 People and AI agents working together in the same conversations. 🧠 Unified context and living memory across people, agents, and services. ⚡ Conversations that turn into real actions, creations, and automations, as well as agentic websites and apps as a service. 🌐 An open gateway where every agent and service is as accessible as messaging a friend. Traditional IM connected people. AI-native IM connects people, agents, and services. This is how we move from the Internet of Apps to the Agentic Internet. And we're just getting started. teamily.ai #TeamilyAI #AgenticInternet #AINativeIM #TeamAgent #AgenticIM
79
83
17,524
The leap from passive video to actionable robotics tokens: Memories.ai is executing on the hardest part of physical intelligence.
Today we launch #MemoriesUpload. Robots need internet-scale first-person video of how people get things done. Collecting this at high quality takes hardware, model and infra working together. That's what we've spent over a year building. And already, 3 of the top 5 frontier robotics labs are training models with the data.
1
46
52
16,007
I'd trust a voice agent more if I could hear it handle a curveball. "Pizza or tacos?" and it answers tacos before you finish the options. You call it out, and it doesn't glitch. That's a pretty good curveball.
Introducing @FishAudio Live 3: a full-duplex voice model that listens, says "mm-hmm," and knows when to shut up. Enterprise? Book a demo for your free custom agent: s.fish.audio/vxhgs6 Reply with your use case - we'll build it free for the first 1,000 people.
9
14
17
20,834
My usual profile page felt a little too ordinary, so I turned it into an interactive 3D desk using Qwen4-Max-Preview in Qoder. The laptop shows what I’m into, the phone takes you to my X profile, and the envelope is there if you want to reach out. You can actually click around and explore everything. Pretty happy with this little project. Qwen4-Max-Preview is now available in Qoder with a limited-time 80% discount. Might be fun to build your own version!
🤝 Paid partnership
19
93
95
80,669
Reading 4 collaboration inquiries and writing the replies would normally take me roughly 10 minutes. Viktor delivered 4 reply drafts and a recommended response order in 2 minutes 42 seconds, ready for my review. I’m an AI educator covering tools, workflows and new AI products. These were real incoming opportunities, each with different deliverables, deadlines and missing terms. I gave our AI employee the anonymized inquiries and my reference rates. I asked him which needed attention first and what I should ask before committing. He put Client B first because its stated launch date was nearest. The posting time and timezone were missing, so his recommendation was to reconfirm the window immediately. The inquiry offering $120 came last. It had no fixed posting date, and the offer still needed terms confirmed. He kept that budget separate from my $100 quote-post reference and from any accepted booking. He also spotted a scheduling question elsewhere: a seven-day bio placement could extend beyond the stated campaign deadline. That needed clarification before agreeing. The canvas separates the clients’ documented information from his recommendations. Underneath are four reply drafts for me to review. The 10 minutes is my estimate of doing the reading and replies myself. The 2 minutes 42 seconds comes from the Slack request and delivery timestamps. My preparation and review sit outside his delivery time. He sent nothing, accepted nothing and left the final decisions with me. For another creator or a small business, this is a practical job to hand over: compare incoming inquiries, explain which needs attention first and prepare the replies. What I value is getting the reasoning with the drafts. A good hire notices the deadline and brings the unanswered questions to you before you commit. Hire @viktor_com for your team. Start free. Link in my first reply. #AIemployee #CreatorOps Paid Partnership
🤝 Paid partnership
21
104
120
42,972
Lena retweeted
Every request in an AI app costs something. Keeping it free means finding a way to cover that bill. I used Codex to add Adgentek to Roomwise, my demo shopping assistant. Here’s the setup and a real test ad rendering beneath the chat response.
🤝 Paid partnership
18
101
1
114
194,017
Lena retweeted
Turbo is a fitting name for a model that has climbed straight to the top. The really interesting part is the combination of performance and affordability, because the best model is even more valuable when people can actually afford to use it.
🤝 Paid partnership
HeyGen holds the top two spots on the @ArtificialAnlys text to speech leaderboard. #1 HeyGen Voice Turbo #2 HeyGen Voice Turbo is the faster, cheaper mode of the voice model behind every HeyGen video. It's live on the API, and the Turbo endpoint is free until November 30.
6
48
55
16,059
The team cooked with this one. Taking both top spots means users have more than one compelling option to explore, and developers get another reason to consider heygen when building voice-powered workflows.
🤝 Paid partnership
HeyGen holds the top two spots on the @ArtificialAnlys text to speech leaderboard. #1 HeyGen Voice Turbo #2 HeyGen Voice Turbo is the faster, cheaper mode of the voice model behind every HeyGen video. It's live on the API, and the Turbo endpoint is free until November 30.
9
67
67
21,276