Building and marketing projects with AI in public. Agency founder after exit, ex-VP ENG at TIER. Dad of two. Optimizing health, sleep, and training.
Budapest, Hungary
Joined May 2009
- Tweets5.8K
- Following871
- Followers2.3K
- Likes21K
Pinned Tweet
Past week in AI news 🧵
- Claude Duet, Anthropic's proactive agent, is now available via a waitlist.
- Dessa, Block's full-duplex speech model was released, trained purely on audio of human conversation.
- OpenAI models are coming to Google Cloud.
- Meta Muse hits 10M downloads in the US in 40 days.
- DeepSeek raised $15 Billion.
- Canva releases Canva Craft, the company's first image generation model.
- NVidia launches two new models: Nemotron 3.5 Super VL (model for always-on agents), Nemotron 3.5 Bolt (agentic intelligence to local NVIDIA systems).
- OpenAI acquired AI-powered drug development platform Orchestra Bio.
- Two new Kimi models were introduced: Kimi K3.1 Pro and Kimi K3.1 Pro Highspeed.
- OpenAI's Decisions API is now available, to help applications and agents make near real-time decisions faster.
This is not an exhaustive list, just the stories that caught my attention. Let me know if I missed anything!
Great to see companies taking action against slop. Some examples from this year:
- LinkedIn introduced a dedicated ‘report AI slop’ button
- YouTube tightened monetisation rules around AI slop
- Substack launched a tool to detect AI-generated writing
- Snapchat to stop recommending wholly AI-generated videos
- Apple started to terminate accounts
- Now Shopify
I just hope this will be enough to take back at least a portion of the internet
Today we unpublished hundreds of apps from the App Store. They came from a small group of high-volume submitters and had no merchant installs.
The App Store exists to solve real problems for merchants. Mass submissions distract from the partners who are doing that work.
We’ll soon be announcing new restrictions that limit the number of apps that can be submitted at once and we’ll keep raising the bar.
Yes, but not just that
The long-term moat of X is being a place for human interactions
If it becomes a bot-to-bot platform, it will lose its value
X should have a zero tolerance policy for AI replies
Like you do one and your account is suspended for life
Replies are a nightmare now and it's never ever been this bad
AI replies are an an existential threat to the future of this platform, because people will simply stop using it if the vast majority of replies are AI bots
Follower count actually feels important again. A few iterations ago it was starting to look more like a vanity metric, now it is a real advantage:
- More people treat your posts as in-network (no out-of-network discount);
- Your posts rank higher for people who follow you back;
- It provides a reliable floor of reach;
- Engaged followers improve the model’s predictions, boosting future out-of-network distribution.
My fav afternoon snacks to reach my daily protein goals when I'm not at home:
- beef or salmon jerky
- dried edemame beans
- roasted chickpeas
- nuts
- rice cake + PB + bananas
- chia pudding
These require little to no prep, and make no mess.
What are your picks?
YC’s latest RFS just dropped, a great indicator for what’s next. These are the latest ideas they’d specifically like to see founders tackle:
- Abundant space access startups
- AI-powered expert workers
- American-made robots/supply chain
- Better-writing LLMs
- Plaid-style government benefits
- Financing physical AI buildout
- Vertical specialized foundation models
- AI tools for city hall
- Data center supply chain solutions
- On-prem AI for air-gapped systems
- Tools to own business AI models
The bottlenecks in AI are moving beyond the models.
We’re excited about startups building the hardware, infra, financial systems, and specialized software to bring AI into the real economy, from factories and data centers to government and space.
ycombinator.com/rfs
P=NP when
Announcing three major scientific discoveries, one each in mathematics (harmonic analysis), computer science (complexity theory), and physics (quantum field theory).
AI starting to help us improve our understanding of the universe! openai.com/index/building-sc…
By not reading and writing, people are essentially outsourcing thinking:
- The algorithm decides what content you consume;
- Short-form and slop degrade your thinking and fry your brain;
- AI formulates your thoughts you share with colleagues, friends, family;
- All of this is so convenient, it is tough to resist.
Sounds like a bleak future, and it is, unless you are able to practice self-control.
What is a neocloud?
A neocloud is an AI-first cloud platform, designed to provide infrastructure for GPU compute and AI workloads.
Unlike traditional cloud platforms offering a wide range of services like databases, auth, analytics etc, neoclouds specialize in GPU compute.