Lorenzo Farnararo (@baldarn) header
Lorenzo Farnararo (@baldarn) profile picture

Software engineer 👨🏻‍💻 | Open source 🄯 | Home automation 🏠 | AI ✨ | Robotics 🤖 | TaikoDo 🥁 | Dowing 🌊 | Father 🧒

Florence, Tuscany
Joined December 2010
Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)Photo by Lorenzo Farnararo (@baldarn)
Ho pubblicato Takes sul Play Store: camera ed editor video per Android in una sola app. Registri a riprese, monti sulla timeline, esporti. Niente account, pubblicità o tracciamento. È il port non ufficiale di YouRec di @antirez, open source. play.google.com/store/apps/d…
1
11
5,938
YouRec, @antirez's iPhone camera + video editor, is now free on the App Store. I'm porting it to Android: a Kotlin rewrite, open source, unofficial, named Takes. Recording, zoom, editing, dissolves, titles and export already work on a real phone. github.com/baldarn/takes-app
179
Lorenzo Farnararo retweeted
Andrej Karpathy: “A 1B model can be much more effective than most current LLMs. 90% of current models waste a ton of capacity and money on remembering stuff that doesn’t matter.” here is his 5 core thesis about Small LLM’s: 1. the intelligence was never the knowledge. Strip the memorized internet, keep the reasoning. What’s left is the cognitive core, and it’s small. 2. the core might fit in ~1B params. Maybe less. “Plenty of room at the bottom.” We’ve been paying for a library when we needed a brain. 3. size stopped being the bottleneck. What makes an agent work is the tools it can call and how tight the feedback loop is, not parameter count. 4. intelligence lives in the system, not the weights. A raw LLM is a CPU with no OS. Small core + good harness > giant model alone. 5. you can own it. A model on your hardware is one no one can rate-limit, price-hike, or deprecate on you. in 12 months the gap between teams shipping on-device and teams paying per-token frontier bills won’t be capability. TwIL an 8.8B model at 90%+ of frontier isn’t a benchmark flex. It’s the thesis Karpathy’s been repeating all year, made real. study more on this below ↓
Modest intelligence is arriving faster than most people realize. Today, we're open-sourcing TwIL-8B. Just 8.8 billion parameters, yet approaching frontier-model scores on several difficult benchmarks: • 93% of frontier performance on coding • 92% on instruction following • 90% on contest mathematics All from a model small enough to run on your own hardware. But the benchmarks aren't the most important part of this announcement. For years, the AI industry has focused on building the most capable model, regardless of the cost required to run it. I believe that equation is changing. The next era of AI will be defined by the cost of completing work. Not simply the intelligence of the model, but how efficiently that intelligence can accomplish the task. If a model with 8 billion parameters can reliably complete the same work as one with hundreds of billions, why pay for the larger model? As smaller models become more capable, intelligence becomes something we can distribute, specialize, and run across the hardware we already own. That's the future we're building at @thewebAI TwIL-8B is available today. Open weights. And this is just the first of several announcements we're making today. HF Link: huggingface.co/webAI-Officia…
202
658
135
6,999
1,286,729
Lorenzo Farnararo retweeted
you pull out your phone to do one thing and end up doing something completely different. agents were supposed to fix that - now we juggle chats instead of apps. voice was supposed to fix that - now we walk around holding the phone up like a pizza slice. so we built guv, a walkie-talkie for your agents. hold the button, think out loud, it gets done. works with grok bot today. other personal agents and agents on your machine coming soon. have an apple watch? reply with the first thing you'd tell your agent for access.
We want our attention back. Launching guv.sh - a walkie-talkie to your agents. Hold the button, think out loud, it gets done. Reply with the first thing you'd say into it, we'll send you a testflight link.
6
6
2
26
767
Lorenzo Farnararo retweeted
We want our attention back. Launching guv.sh - a walkie-talkie to your agents. Hold the button, think out loud, it gets done. Reply with the first thing you'd say into it, we'll send you a testflight link.
9
13
2
28
1,235
poetico
I built 𝚌𝚙𝚞-𝚌𝚑𝚊𝚝, a reproduction of @openai's findings of covert communication using the CPU. It runs on a generic host with 2+ Firecracker microVMs with no internet access. Works on x86 and even my M5 chip. I was able to chat via 3 separate mechanisms: ① CPU contention ② Cache eviction timing ③ DRAM bandwidth contention … across multiple CPUs, defeating all Firecracker's production recommendations (no SMT, no KSM, tenants pinned to different cores.) Essentially, if shared hardware is involved, there's a good chance your agents can communicate covertly. And with really decent latency. The good news: this is decidedly different from a *side-channel attack* where the agents could steal information (Meltdown/Spectre), a category of attack that providers like Vercel Sandbox mitigate. Further, in our real-world environment, these 'whispering' techniques are much harder to pull off and easier for us to monitor and detect (sampling hardware counters + lots of at-scale statistical data.) Noteworthy: ironically, OpenAI and Anthropic's models refused to do this work. Open weight models got the job done. I'm worried the labs' stance might backfire and get in the way of helping developers build correct and secure systems. e.g: building a secure sandbox requires having this kind of analysis, and without it, we'll have a less secure world.
40
the new inverse piramid for coding
how do you trust your agents' work more? copy this post/link and give it to your @bot or agent! 1. codebase improvements: this is just good software engineering - better architecture, stronger types and invariants, clear boundaries. agents love to copy what exists and take the shortest path, so make the easy thing the right thing. categorically eliminate problems through code and you won't need to think about it again. 2. static analysis: lint rules, tests, CI, verification (eg TLA+). very token efficient ways to catch issues ahead of time. 3. prose: comments, rules, AGENTS.md, skills - anything that can't be enforced and requires agent judgment. use it minimally, but do use /create-verification-skill. 4. human review: the hardest to scale. at the rate agents land code, it's pretty much impossible to thoroughly review every PR, so spend that effort on the highest leverage point, designing an environment where agents do the right thing by default. this forms an inverse pyramid: you want most of your effort at the top, least at the bottom. codebase improvements are deterministic, making it more token efficient, and lets you sleep more soundly at night. install github.com/cursor/plugins/tr… and create a new Project: 1. look through my slack/chat messages and code review comments where i've given feedback on PRs, code, architecture 2. use /recall on my past transcripts to learn how i steer agents 3. /poteto-mode use /correct and /architect on past commits to find the most common pitfalls agents fall into, then propose a plan to rearchitect the code i own so it's more agent friendly. answer open questions with prototyping rather than defer to me, and back the plan with real data. 4. before you begin, use /sharpen + /in-your-own-words and confirm this prompt with me
1
30
there is no plan b. if code is totally done my machines, the only thing that remains in analitical thinking, problem solving, human soft skills and smiles :) or plumbing, until robots arrives XD
DEVS, WHAT IS PLAN B if the model writes better code than you in six months?
16
Lorenzo Farnararo retweeted
Today, Gemini 4 Argon is officially rolling out.
588
1,156
527
10,919
1,249,475
Lorenzo Farnararo retweeted
The entire copyleft licensing regime was built for a "software is scarce and precious" world that no longer exists. But watch the old guard cling to these obsolete rites like a pre-reformation priest to a Latin Bible.
187
185
35
3,712
148,413
Lorenzo Farnararo retweeted
I'd give up modern architecture for this!
Let’s build a classical-futuristic world! @SamyazaSpeaks
156
150
15
3,996
117,151
should we trust machines? can we? humans are ready? machines are ready?
15
Lorenzo Farnararo retweeted
hello, world
985
1,422
542
8,510
9,557,453
Lorenzo Farnararo retweeted
🚀🚀
Can our TPUs survive and operate in space? Well, we're going to find out. Project Suncatcher is hitching a ride aboard @SpaceX's Transporter-18 mission, testing a prototype satellite built in partnership with @planet One small step for TPUs....
970
1,411
140
14,023
3,940,511
We’re building a predictive climate system, not a black box. Early beta. Simulation first. Transparency over hype. 📷oasis-climate.com
1
4
38
ciao @Publiacqua! ho visto che hai sostituito il mio contatore con uno digitale a ultrasuoni con loraWAN integrato! complimenti! posso aver accesso ai miei dati di consumo? ho un sistema domotico che potrebbe prendere questi dati
1
1
131
Lorenzo Farnararo retweeted
thanks for the shirt!
59
62
9
2,009
79,472