John Werũ II (@werucome) header
John Werũ II (@werucome) profile picture

Two roads diverged in a wood, and I— I took the one less traveled by, And that has made all the difference.

East Africa
Joined May 2011
Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)Photo by John Werũ II (@werucome)
📢 #WeAreHiring! The High Commission of Canada in Kenya is looking for a Trade Program Coordinator to join our team. Interested? Learn more about the position and submit your application by 🗓️November 18, 2026. #IkoKazi 🔗 staffing-les.international.g…
1
53
3
156
47,286
John Werũ II retweeted
If you are concerned about how to build and scale apps to millions in downloads and revenue, then this is a must watch! Alex is great at growth.
I get A LOT of questions about how to build and scale successful apps. So I spent 6+ hours filming a video sharing almost everything I've learned along the way. Within just 24 hours, it became my most viewed video ever, surpassing all of my vlogs. I hope it inspires some of you to go out and build some incredible companies :)
1
2
90
John Werũ II retweeted
MAGENTA hiring, #IkoKaziKE #JobsearchKE - Finance & Operations Associate - Nairobi, Kenya / Lisbon, Portugal - USD 72.22 per day Apply: Via application link in advert Deadline: 12th December 2026
3
72
275
27,063
John Werũ II retweeted
MIT has an entire course on the stuff you should know at your first engineering job shell, git, debugging, profiling, shipping code the 2026 version even has agentic coding. all free. link: missing.csail.mit.edu/
36
344
17
4,725
153,254
John Werũ II retweeted
8.2 TB downloaded in the last 31 minutes.
100
223
9
7,534
1,486,325
John Werũ II retweeted
Mental models that runs almost 93% of your life.
212
6,652
173
45,209
2,667,782
John Werũ II retweeted
Mental models that runs almost 93% of engineering
44
1,437
27
8,866
316,509
John Werũ II retweeted
The smartest download this week is Nvidia's free AI. Nemotron 3.5 Bolt E4B is a 4GB file that runs on almost any laptop. It writes, answers questions and drafts emails with Wi-Fi off and nothing leaves your machine. Grab it, while it's still up.
91
296
18
3,268
273,446
John Werũ II retweeted
I’m convinced mathematics is taught wrong. Its beauty is buried under technical jargon and lifeless formulas when it should be clear, visual and practical. My mission is to get you fluent in the math that will still be useful in 50 years. Start here: thepalindrome.org/p/the-pali…
7
49
2
427
13,403
🚀 Applications are open for the @DigitalAfrica__ Seed Fund, offering investment to African tech startups ready to scale. Digital Africa plans to invest €300,000 to €2 million in promising technology companies across nearly 20 priority African markets. The fund is targeting around 30 startups, with particular interest in ecosystems that have historically attracted less venture capital. Priority areas include: 🤖 Artificial Intelligence 💳 Fintech 🏥 HealthTech 🌱 ClimateTech 🛰️ Space Technology 💻 Digital Infrastructure Applicants should be at MVP or traction stage and demonstrate a strong business model, capable founding team, regional growth potential and measurable impact. Selected startups will receive investment alongside structured support designed to help them strengthen execution and prepare for subsequent funding rounds. 🔗 Find Details Here: menterprise.africa/opportuni…
2
22
2
51
2,576
TO WHOM IT MAY CONCERN 🌍 He got 4 offers without applying to a single job listing. Here's how: ✅ Catch "we're hiring" posts less than 1 day old and email the poster directly ✅ Follow seed funding news and email the founders ✅ Send 20 short emails a day, NOT written with AI You still need the skills. This only gets you SEEN. Be STRATEGIC!
Bro found a loophole in the job market 💀
7
36
4,316
John Werũ II retweeted
Read the textbooks behind robot brains for free 1. The math mml-book.github.io 2. Probability probml.github.io/pml-book/bo… 3. Deep learning, by Goodfellow, Bengio & Courville deeplearningbook.org 4. Computer vision, from MIT visionbook.mit.edu 5. State estimation (how a robot knows where it is) asrl.utias.utoronto.ca/~tdb/… 6. Motion planning lavalle.pl/planning 7. Control fbswiki.org 8. Reinforcement learning, by Sutton & Barto incompleteideas.net/book/the… 9. Decision making algorithmsbook.com 10. Robot learning, by the Hugging Face LeRobot team huggingface.co/spaces/lerobo… P.S. #8 is written by the two researchers who won the 2024 Turing Award for reinforcement learning, the method robots now use to teach themselves to walk. the full PDF is on their own site which order to read them in is the free 12-week path in the article below
Stanford charges about $70,000 a year to teach robotics. NVIDIA, Google DeepMind and Hugging Face teach it for $0. i put their free courses in order: 12 weeks, laptop first, $230 robot arm last. save it for the weekend
4
116
2
644
19,369
John Werũ II retweeted
THIS IS AN ACTUAL F*CKING GOLDMINE someone turned Opus 5.5 into your own motion design studio with 637 animation techniques every solo designer can steal the workflow all organized into a guide and agent skill for Claude Code and Codex the techniques to steal: → reveal products with animated masks → animate headlines with kinetic typography → guide viewers through your UI with clicks and highlights → turn metrics into moving charts and counters → explain agent systems with animated paths and diagrams → add depth with parallax, camera moves and 3D card flips → combine effects into product demos and launch videos the useful part: effect cards include timing, easing, parameters, examples and prompts your agent gets specific references to build from the workflow: describe the scene → choose the technique → adapt the example → render → inspect start with your next product launch: one screenshot → one key interaction → a 10-second demo save this, then read the full guide for Opus 5.5 Motion Engineering ↓
Opus 5.5 can run a whole motion design studio Sell a 20-second launch clip built from real product screens Copy the motion engineering prompts and delivery process Build your offer before the next product launch ↓
14
23
150
12,999
John Werũ II retweeted
Atlas is now available to everyone!! I’m very excited to see what users can build with our newest world model! 🤩🌎
We’re opening up access to Atlas to everyone today. We’ve built apps and API endpoints to help you use Atlas any way you can imagine: bring your ideas for building a 3D world and come away with keyframes, video clips, or 3D Gaussian splats.
108
308
51
2,720
179,211
John Werũ II retweeted
40 people will get a fully funded scholarship to learn how to test AI models. It's called the International Programme on AI Evaluation: Capabilities and Safety, run by ValgrAI. What you get: • Full tuition paid by a scholarship • Hands-on training in evaluating what AI can do and where it's unsafe • A mentored team capstone project • A ValgrAI Expert Diploma when you finish What you don't need: • To move anywhere, it's mostly online • To quit your job, it's 12–16 hours a week • To be in a specific country, it's open worldwide The condition: you're a grad or PhD student, an early-career researcher or a professional with some technical skills. Runs Feb to Jun 2027. Deadline: Dec 11, 2026. AI labs are hiring people who can measure models not just build them. Apply here: ai-evaluation.org/apply Bookmark it and send it to the friend who wants to break into AI safety.
8
26
1
204
7,764
John Werũ II retweeted
THIS IS AN ACTUAL F*CKING GOLDMINE Anthropic just shared one of the best tricks for Claude Code. The idea: run Sonnet, Haiku 5.5 and Opus 5.5 at once, each one doing only the job it's best at. Better results and way fewer tokens, because every task lands on the right model. You just launch Claude Code like this: claude --model sonnet --advisor opus here's the split: → Sonnet 5.5 runs the session, writes the code and applies the fixes → Haiku 5.5 swarms the repo in parallel: search, grep and diff audits → Opus 5.5 stays on call as advisor for the big calls and signs off on the diff → you stop burning Opus context on work that doesn't need it Opus only steps in at three moments: → before a plan locks → when the same error shows up twice → before calling it done Sonnet builds. Haiku swarms. Opus reviews. JEV engineering is the same move one layer down: the forks that need no thinker (which file, which tool, retry or stop) go to JEV in milliseconds, and the big models only see the ones that split the full tree > Sonnet 5.5 runs the main session and coordinates the work > search maps the repo on Haiku 5.5, read-only > grep finds call sites on Haiku 5.5 in its own context > audit checks the diff on Haiku 5.5 in a throwaway worktree > Opus 5.5 on call as advisor, reviews the synthesized diff before sign-off > JEV routes micro-forks in milliseconds paste the tree and this prompt into Claude Code ↓ "Reconfigure my Claude Code setup around this tree: 1. Run the main session on Sonnet 5.5 with Opus 5.5 as advisor. 2. Create three Haiku 5.5 subagents in .claude/agents: search (maps the repo), grep (finds call sites), audit (checks the diff). 3. Set isolation: worktree on audit so its checks never touch my working branch. 4. Add one rule to CLAUDE.md: delegate every repo search and test run to the Haiku 5.5 subagents to keep the main context small. 5. Route fast deterministic forks (tool choice, path selection, retries) through JEV. Show me every change as a diff first. No edits until I say go." ↳ code.claude.com/docs/en/sub-… save it, it'll come in handy.
this is pure f*cking treasure... Anthropic just released the most dangerous model they've ever shipped everything you need is inside: what it built, the real costs, when it's a waste, and 5 prompts that build anything read before you burn your limits:
18
44
2
371
49,592
John Werũ II retweeted
Son Dakika. Robotik öğrenmenin en pahalı kısmı ortadan kalktı. Bir geliştirici, Fable 5.2 ile tamamen interaktif bir “Robotik Mühendisliği” eğitim sitesi açtı. Artık GPIO pinini öğrenmek için 500$+ başlangıç kiti almana gerek yok: • 3D ortamda 8 gerçek kart: Arduino Uno, Raspberry Pi 4 ve diğerleri • Hiper gerçekçi grafikler, gerçek hayattaki hareketlerle birebir • Parça parça inşa et: Çipleri, konektörleri ve başlıkları yerleştirip lehimleyebiliyorsun • Gerçek devreleri bağla: LED yak, sensör oku, motor çalıştır En dikkat çekici detay: Tüm 3D modelleri Fable 5.2 kodla üretti. Hazır 3D dosya veya fotoğraf kullanılmamış. Ücretsiz ve kayıt gerektirmiyor. robotics-lab-app.vercel.app/… Bu tarz projeler çoğalırsa donanım eğitimi tamamen değişir.
11
525
15
3,244
179,421
John Werũ II retweeted
"Kenya's economy could be more than 20 times larger by 2035.." That's an INSANE projection. People waste time talking about a lot of things.. when changing a single thing could change.. literally everything. For the better, massively.
When is someone going to focus on this as their platform and nothing else? It seems like it should be a pretty easy and winning electoral pitch. 🇰🇪🗳️ 2035 is only nine years away. "Kenya's economy could be more than 20 times larger by 2035 if it introduced the sort of anti-corruption reforms seen in countries like Ghana and Senegal, and add 7.7 million more and better jobs."
5
17
1
70
2,779
John Werũ II retweeted
Before applying for an AI governance role, spend 20 minutes on the company`s public pages. Look for an AI policy or responsible AI page, any AI features in their product, and whether they sell in the EU. Check recent news for an AI complaint or lawsuit too. Then mention one specific thing you found in your cover letter (or you can wait til first interview). Hiring managers notice when someone has looked. BE PROACTIVE. Comment research and I`ll DM you the checklist I`d use 🙂
87
44
395
13,835
John Werũ II retweeted
this Anthropic engineer's (Addy Osmani) paper is an actual goldmine for anyone building with AI... his point: the better your agents get, the easier it is to stop questioning them a 9-step blueprint for keeping your edge while agents handle the work before trusting the output: → write your hypothesis before prompting → ask what would make the answer wrong → inspect the actual test output before the summary → surface the assumptions and decisions the agent made → explain what you're shipping well enough to defend it i packaged the practices into a reusable review prompt: have Claude list what each check proved, what it missed, and then quiz you on the changed behavior 6 questions. 9 practices. one workflow for staying in control of what you build save this prompt => send it to your Opus 5.5, thank me later then read the full guide for Opus 5.5 Harness Engineering ↓
Stop letting unfinished Opus 5.5 runs eat your budget Set up the harness before your next long task Save progress, verify results, reuse the workflow The full 7-layer setup below ↓
36
18
183
7,288