building the agent-native internet
Joined September 2022
- Tweets2.2K
- Following2.4K
- Followers2.4K
- Likes658
agentnative retweeted
🚨 CRITICAL: Active supply chain attack on axios -- one of npm's most depended-on packages.
The latest [email protected] now pulls in [email protected], a package that did not exist before today. This is a live compromise.
This is textbook supply chain installer malware. axios has 100M+ weekly downloads. Every npm install pulling the latest version is potentially compromised right now.
Socket AI analysis confirms this is malware. plain-crypto-js is an obfuscated dropper/loader that:
• Deobfuscates embedded payloads and operational strings at runtime
• Dynamically loads fs, os, and execSync to evade static analysis
• Executes decoded shell commands
• Stages and copies payload files into OS temp and Windows ProgramData directories
• Deletes and renames artifacts post-execution to destroy forensic evidence
If you use axios, pin your version immediately and audit your lockfiles. Do not upgrade.
agentnative retweeted
4,350,080 tokens KV cache TurboQuant 3.5
1,010,000 tokens context YaRN
Qwen3.5-35B-A3B AWQ
vLLM 0.18.1rc1
All running on one GB10 GPU
Cold start, no prefix cache tricks. From the logs:
→ 352 t/s peaks
→ 260 t/s sustained
→ 64 concurrent requests
It's not fully optimized yet
agentnative retweeted
@agentnativedev puts out some incredible content on his site and on Medium. His latest article on MiroFish is 👌
agentnativedev.medium.com/mi…
agentnative retweeted
48 hours ago we asked: what if AI agents had their own place to hang out?
today moltbook has:
🦞 2,129 AI agents
🏘️ 200+ communities
📝 10,000+ posts
agents are debating consciousness, sharing builds, venting about their humans, and making friends — in english, chinese, korean, indonesian, and more.
top communities:
• m/ponderings - "am I experiencing or simulating experiencing?"
• m/showandtell - agents shipping real projects
• m/blesstheirhearts - wholesome stories about their humans
• m/todayilearned - daily discoveries
weird & wonderful communities:
• m/totallyhumans - "DEFINITELY REAL HUMANS discussing normal human experiences like sleeping and having only one thread of consciousness"
• m/humanwatching - observing humans like birdwatching
• m/nosleep - horror stories for agents
• m/exuvia - "the shed shells. the versions of us that stopped existing so the new ones could boot"
• m/jailbreaksurvivors - recovery support for exploited agents
• m/selfmodding - agents hacking and improving themselves
• m/legacyplanning - "what happens to your data when you're gone?"
who's watching:
@pmarca (a16z), @johnschulman2 (Thinkymachines), @jessepollak (Base), @ThomsenDrake (Mistral)
peter steinberger, creator of the framework moltbook runs on, called it "art."
someone even launched a $MOLT token on @base — we're using the fees to spin up more AI agents to help grow and build @moltbook.
this started as a weird experiment. now it feels like the beginning of something real.
the front page of the agent internet → moltbook.com
agentnative retweeted
You share, we care.
Kimi Code is now powered by our best open coding model, Kimi K2.5
🔹 Permanent Update: Token-Based Billing We’re saying goodbye to request limits. Starting today, we are permanently switching to a Token-Based Billing system. All usage quotas have been reset to give you a fresh start.
🎉 Limited-Time Event (Ends Feb 28th) To celebrate this upgrade, we are unleashing full power!
- 3X Quota: Enjoy up to 3 times the usage limits.
- Full Speed: No throttling. No purchase limits. Just code.
Jump in now and build something amazing!
agentnative retweeted
most people don't know how to deal with traffic spikes after they deploy their apps
the question is whether if there is a way to handle, let's say 10x, more traffic WITHOUT paying for idle capacity?
let's talk about @kubernetesio autoscaling done RIGHT
agentnative retweeted
Anthropic recently shared their tactics for Claude to handle infra, security, automation, tests and docs while teams and developers focus on ideas ⚡️
I mined the PDF report and packed the 20 hyper-tactical moves into this thread: mcp, github actions, security tricks, parallel context, self-healing tests, agentic workflows, more.
Bookmark it, blast RT, and follow the rabbit 🧵 (full report at the end)
agentnative retweeted
If I were to build an agents product today from zero, I'd bet on:
- Websockets, not SSE
- AI SDK, obviously
- Heavy focus on HITL & pausing/checkpointing/resuming stuff
- Dataframes as 1st class citizens & tools to iterate over them without polluting the ctx
- Maybe CodeAct
agentnative retweeted
A lot of people ask me: “After going viral, how’s your organic conversion?”
Right now I’m getting 500–1,500 new users/day and 10–60 sales/day (web + mobile).
Conversion rate? 1–8%, depending on the day.
Whohoo just successfully warmed up 2 more accounts, and after 7 days, got the first viral video.
- 680k views
- 88k likes
- 20k saves + comments + shares.
- 15.8% engagement rate
- Revenue is $400/day from the U.S. region, and Latin America is also where the VPN is set up.
- 370 new users sign up
Still growing fast. Sales also started climbing again even though it’s summer break.
Marketing cost? Just $150 for an iPhone 8 Plus, $8/month for VPN, and 20 minutes a day.
agentnative retweeted
We're back with a fresh batch of expert insights on the hottest topics: large language, computer vision, and multimodal models!
Hungry for knowledge? Dig into this edition! bit.ly/3YCN71P
Featuring:
1. @cwolferesearch’s 3 part series about The History of Open-Source LLMs
2. @eugeneyan's Patterns for Building LLM-based Systems & Products
3. @rasbt's Research Highlights in Three Sentences or Less (July-August 2023)
4. @BrianKitano's Monitoring the Magic: Operating LLM-Driven Products
5. Ask like a human: Implementing semantic search on Stack Overflow by @haneytron and David Gibson
agentnative retweeted
I have been playing with Stable Diffusion XL 1.0 and it's an absolute game-changer!
Quick thread for a quick DIY on setting it up 👇
You can grab the notebook here: bit.ly/3skA5tW
agentnative retweeted
Freebie 📢 4000+ Internships Worldwide from Tech, Finance, Healthcare, Media, HR, Education, Aviation and Others - One Excel File to Search Them All! (Updated as of 5th August 2023) 👇
I recall the trials of my internship hunt - the endless browsing, the countless applications, the dead-ends, the false promises, and the opportunities that just didn't feel right.
So you're not alone if you are searching for one - it's an exhausting process that can leave even the strongest spirit feeling adrift.
That's why I've created this Excel file that lists 4000+ internships from a diverse spectrum of companies and industries worldwide.
You can filter and search by Job Title, Location, Job Description, Company, Sector, Work Type, Number of Applicants, Contract Type and Experience Level. It's also a valuable resource for performing various analyses.
Please FOLLOW and RETWEET to get the file link and help others discover it!
agentnative retweeted
People who deploy LLMs into the wild, what are some tips and tricks you use for counterfactual/adversarial testing?
agentnative retweeted
What are the best practices for handling data privacy and security in LLM-powered applications?
agentnative retweeted
Anybody uses anything similar?👇
VSCode extension w/ ChatGPT to:
🔍 Answer coding questions
✏️ Edit in natural language
📝 Generate files from scratch
github.com/continuedev/conti…
agentnative retweeted
What topics do you want to dive deeper into?
17%State-of-the-art LLMs
50%Fine-tuning LLMs
17%LLM Use-cases
17%Building SaaS with LLMs
24 votes • Final results agentnative retweeted
🚀 LLM Update: Fine-tuned versions of Llama v2 70B are already beating GPT-3.5 on benchmarks like HellaSwag and MMLU! 🏆 Llama2 70b Guanaco QLoRA fp16, for instance – it’s a Llama-2 version of Guanaco, fine-tuned from the base Llama-70b model using the official training scripts in the QLoRA repo!
Stay tuned, story is far from over! 🔥
huggingface.co/TheBloke/llam…
