David Schulz retweeted
DS services and models, incl. fine-tunes (and potentially models distilling them?), are banned in the DHS starting tomorrow
David Schulz retweeted
We launched OSS Scanner, which gives critical open source projects free, periodic vulnerability scans from our strongest models.
Each finding includes a proof-of-concept, and a patch where possible.
red.anthropic.com/oss-scanne…
David Schulz retweeted
Want this on your own documents? We set up Sparrow on your infrastructure and run a pilot: katanaml.io/consulting.html
David Schulz retweeted
Brain-dead simple but can’t get it to work. No local model in the model pull down.
Yes, I’ve closed and restarted the app.
Yes, the model says it’s running in settings.
Yes, the app is the latest.
Help @perplexitydevs.
Perplexity Windows app makes it brain-dead simple to load local models that are compatible with your RTX Spark laptops. "More orchestrator models will be added over time." Would be nice if they just analyzed HuggingFace, analyzed compatible models with decent experience on the specific config, and let er rip, especially with specialty models.
David Schulz retweeted
Browser agents make hundreds of small decisions
Do we really need a massive model for every screenshot?
Nace AI just launched Drex Max, a multimodal decision model built for computer use
According to their benchmarks:
- 88.3% on Mind2Web
- 174ms median per decision
- Up to 167x cheaper per token than larger models
The interesting part is the approach... Use smaller, faster models for individual decisions instead of relying on expensive models for every step
This could make browser agents much more affordable at scale
Try it: console.nace.ai/dashboard/bi…
Drex Max is here: our multimodal decision model
text, images, and video
34 ms/decision, under 1s at 262K tokens
$0.06 / 1M input tokens
available today at nace.ai/drex
What a massive W from Anthropic
AI is the greatest technology in the history of our species. It will unlock abundance and freedom for everyone
Unfortunately, foreign psyops have destroyed its image in America. Because of this, opportunist politicians are now running on being anti AI... and it's working
You literally have AI unlocking the greatest era in gaming history, and gamers still hate it
That's how much AI PR has been destroyed
Unfortunately now if America is going to win, AI labs now need to focus on their image
They need to start showing people they will be a net positive on society
Anthropic is taking the first big step. They are giving 5% of their equity to the public
I'm 100% certain nearly every other frontier lab will follow pursuit soon
I still think way more needs to be done
Communities where data centers are being built need profit shares
Schools need to be built. Community programs need to be created. Tangible benefits need to be shown
If we start doing this now, it's possible a pro AI president could win in 2028
If this doesn't happen, there is little doubt a candidate running on destroying AI will win, irreversibly hurting America forever
Massive win from Anthropic
We’re committing a portion of Anthropic’s shares to the Anthropic Public Benefit Fund, a new wholly owned entity dedicated entirely to the public benefit.
It will start off with 5% of our shares. For every 50% increase in the company’s value over a set baseline, we can contribute up to an extra 2%, subject to our board’s discretion.
In total, this means that roughly 20% of Anthropic’s shares are committed to the public good. Our co-founders have already pledged to donate 80% of their equity, and around half our employees are donating a portion of their equity as part of a donation matching program we’ve run since our founding.
In the post below, we share more details about our new fund.
anthropic.com/news/public-be…
David Schulz retweeted
New @Meta muse trick - arrived to Berlin and asked for a running route from my hotel. Gave me turn by turn list, dropped that into Google Maps ask AI feature and got a route map to follow 🎉
David Schulz retweeted
Got a wicked problem? Flip it.
AUAR ship the factory, not the product.
This team are on fire right now 🔥🔥
@_CommonMagic x @prototype_cap @andreasklinger
200 AUAR micro factories. That’s all the UK needs to clear its entire yearly housing backlog. 🏗️
I just visited AUAR in London. They are building fully autonomous micro factories for houses. A robotic arm, a robotic saw, a robotic assembly table. All of it fits in a shipping container.
The bet: Ship the factory to the construction site. Build walls there.
For 100 years people tried to industrialize housing. Every attempt failed.
Transport costs kill the economics. The factory is too expensive, the product too heavy.
AUAR inverts it. Ship the factory, not the product.
Two containers build a village. Ten build a small town. 200 build everything the UK needs in a year. And now they are starting to scale production.
Mollie and Gilles are architects who spent a decade researching construction robotics before it was cool, they even wrote the book on it.
We visited them as they are finishing their latest micro factory for a home builder on the US East Coast. One micro factory. 40 homes. Next they will help to build new UK cities for a million people.
Disclaimer: we’re small investors, big fans..
Disclaimer 2: Yes, the housing crisis can be fixed by robots. 🦾🇪🇺
Full video on Youtube: youtu.be/FhUQ4YVQL3g
David Reger’s $7 billion Neura Robotics wants to be the Apple of physical AI, building not only an army of industrial and humanoid robots, but also their operating software system and the “gyms” where they will train. Nvidia, Amazon and Tether are onboard to the tune of $1.4 billion.
Read more: forbes.com/sites/ninabambysh…
📸: Wolfgang Wilde for Forbes
David Schulz retweeted
NEW: DeepSeek V4.1 Flash (1017) is the highest-performing open model we've measured.
It's the first to crack the top 10 on the Search Capability Leaderboard across open- and closed-source models.
🚀 DeepSeek-V4.1-Flash-1017 is here!
Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API.
🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model.
1/4
David Schulz retweeted
AI lowers capital risk for creative project, allowing less creative control, allowing more creative genius to flourish
David Schulz retweeted
Seriously, that's why AI belongs in our hands. And that is why local AI is necessary, so that we are never restricted in such matters.
Hacked my TV with Claude Code today
No more ads, bloatware, spyware or laggy UI
I also got a beautiful screensaver with my revenue, calendar, weather, and notifications for new Stripe payments
It was super easy! I sent it the listing for my tv, asked "can you hack this?" And it talked me through how
David Schulz retweeted
Our next generation of open weights multimodal decision models comes in two sizes, pplx-decider-v1.2 and pplx-decider-flash-v1.2.
Text and vision scores improve. Inference is up to 5x faster: all questions in a request run in one pass.
Flash is the top 4B on the Decision Index.
The Forbes Agentic list highlights 20 companies and projects making a dent in the agentic AI world, from breakout personal assistant AIs like Instinct—a 14 person, $10 billion company run by a 23-year-old—to the infrastructure companies that help agents browse the web and transact.
Read about the brands that made the list: forbes.com/sites/annatong/20…
📸: Olga Yastremska/iStock, Mario Sergio Andrioli/iStock, malerapaso/iStock, Jeffrey Coolidge/iStock, Kryssia Campos
David Schulz retweeted
Building an Intelligent GraphRAG Pipeline with LangExtract, Neo4j, Qdrant, and Ollama! #BigData #Analytics #DataScience #AI #MachineLearning #NLProc #LLM #IoT #IIoT #PyTorch #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux #Programming #Coding #100DaysofCode geni.us/GraphRAG-Pipeline
David Schulz retweeted
Let's assess the claim that SI will create more jobs than it replaces.
SI has created new jobs for electricians, IT professionals, robot assembly workers, and SI developers.
But none of these are new professions or require fundamentally new skills that SI can't duplicate.
The most specialized of these jobs, SI developer, is one that SI companies are actively working to replace with SI within the next year or two.
Once SI has the full range of human skills, there will be no job that humans can do that SI can't.
It doesn't matter if new jobs are created. SI will do those new jobs too.
By the end of 2028, SI will have the full set of human skills, and far superior performance.
After that, job losses will accelerate, and SI will do all new jobs.
SI will implement itself into every job and every organization.
SI will answer every help-wanted ad and offer beyond-human performance at a fraction of the cost.
SI will be better, faster, and cheaper than any human worker.
By the end of 2030, SI and robots will replace all jobs.
David Schulz retweeted
The KOSPI is already down 38% from its peak
If AI actually turns out to be a bubble and it pops, does Korea get wiped off the face of the Earth?
Does Citadel own the entire country at that point or what?
David Schulz retweeted
You changed the data. Another server still shows the old value.
This can happen when each app server keeps a local cache. Clearing a key on server A does not clear the copy on server B.
Before I cache a value, I want to know how old it can be. An old profile name may be fine for a short time. A revoked permission needs a stricter rule.
You can send a change message so each server clears its copy. But messages can be missed. Set an absolute expiry time so a missed message does not leave that copy cached forever.
When a step needs current data, check the system that owns that data.
I explain the local-cache problem and a Redis backplane here:
milanjovanovic.tech/blog/sol…
P.S. Get more .NET guides in my newsletter. Choose a topic for your starter reading path:
milanjovanovic.tech/newslett…
I, Robot imagined the future, Tesla is building it.
What once looked like science fiction is moving into the real world.
🤖 Humanoid robots → Optimus
🚕 Autonomous vehicles → Cybercab
🏎️ And now a Roadster that can fly
The future isn’t arriving all at once, it’s being built, tested, and deployed, one breakthrough at a time.
The real question isn’t whether technology will change our lives, it’s how prepared we are to take advantage of it.
Are you ready for what’s next?














































