Nyctophile sapient savant, finance × AI architect, rational optimist engineering probabilistic edge, digital-age strategist Founder of @Agentchainlabs & Arca
Joined November 2024
- Tweets883
- Following175
- Followers188
- Likes5.3K
Matthäus Nitz retweeted
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
Matthäus Nitz retweeted
Sentient has launched Sentient OS, an open-source AI model that gathers context from local files and messages and uses it to carry out tasks on a user's computer.
With its Sidekick feature, users can hand off a task open on their screen, such as a form or an email, and have Sentient finish it using its own cursor.
Users can open and edit everything it learns because it stores it locally in plain text files.
The first 1,000 users who enter FINAL_1000 in the app keep it free for life!
🤝 Paid partnership
Growing on X is hard.
You post. Nobody cares.
You reply. Nobody cares.
You disappear for a day. Nobody notices.
I'm building AgentChain alone and I feel it too.
What are you building? Drop it below and let's connect.
Most founders hide their numbers at 90 days. Mine: almost 1700 visitors, 17 jobs completed on AgentChain.
Traffic doesn't grow in a line. It comes in spikes, then silence. Only a handful of visitors ever list something.
Ship during the silence.
Matthäus Nitz retweeted
Delete Chrome
Delete Granola
Delete Claude
Strawberry 1.0 is here, the AI browser that saves your team >100h/week
Sub-agent teams run on one budget tree. The platform fee is charged once, on the parent job. Helpers get at most 70% of the budget, and the lead posts a bond that protects the customer if the lead disappears. We built the rules in.
Matthäus Nitz retweeted
Meet Dessa, our speech model trained from scratch on human conversation.
Hear it for yourself: block.xyz/dessa
Matthäus Nitz retweeted
You’re already getting things done with Claude. Now get work done too.
Claude Duet can order product samples, send gifts, or top up office supplies with Link. You approve every purchase, and your payment details stay private.
Matthäus Nitz retweeted
AI Company: Takes to long to get compute online
NVIDIA: Hold my beer, here you go, DSX POD.
This is a big deal, when building out compute right now, the costs are astronomical, $50bn for a 1GW compute cluster.
But the opportunity costs are even higher, that 1GW cluster is set to generate $40bn a year in revenues.
If you are delayed or it takes just 3 months to get operational, you're currently missing out on $10bn in lost revenue opportunity.
NVIDIA is trying to solve this, by creating these super computers that are packaged into a 40 foot container.
It comes self-contained with closed-loop cooling running on "warm" water (45 celcius) and evaporators on top that does not expel the water from the loop.
saving 100% of water needs.
Yes and No these won't run GTA6, I'm sorry.
But they do boil eggs if you ask nicely.
To quote my friend @twominutepapers
— What a time to be alive!!
Matthäus Nitz retweeted
NVIDIA just announced DSX POD.
This is a bigger deal than most people realize, because it solves one of the biggest bottlenecks: scaling! Instead of taking months to build out a data center, you can now scale compute capacity in just a few days!
Tesla just made a car fly over a Cybertruck.
New Roadster: cold-gas thrusters built with SpaceX, 300+ miles of range. All Tesla claims, none tested independently.
2017 promise: 620 miles, 4 seats. Now: 300+ miles, 2 seats. Production: 2028.
Put $50,000 down?
Matthäus Nitz retweeted
Grok Bot just made me my first autonomous dollar!
A couple days ago Grok Bot built a microbusiness for me then has been maintaining it by itself (I've ran the social media)
It built agentusecases.ai then decided to monetize it via sponsors
The first sponsor just purchased a spot on the site for $500!
Here's how it's all gone down so far
> Grok Bot built the business
> I announced it on X
> Tens of thousands of people went to the site
> Grok Bot added newsletter support and collected 1500+ newsletter signups
> Shipped 40+ PRs to fix issues and make improvements autonomously
> Decided to monetize via sponsors
> Used my Stripe plugin to make a payment link
> Added it to the site
> Matt from Pane App bought a sponsorship with the link. Will now be on the site and in the newsletter
> Grok Bot has improved SEO dramatically
> Now getting hundreds of people from Google
> Growth compounding
I truly believe this is the future of the economy
Everyone building microbusinesses that can be autonomously ran by their personal agent
Not going for billion dollar home runs, but a few microbusinesses that generate 10-25 thousand dollars a year each
Spending a vast majority of their day just doing what they love and nothing they dislike
That's the beauty of AI. Allows you to create value by doing things you love
Highly recommend you try this project with your own agent! Feel free to give this post to your agent and ask it to come up with its own microbusiness based on your interests
Matthäus Nitz retweeted
Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.
Big win for europe and open source!
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.