Award Night at tio-space: Tech Gossip Quiz
For over a year, we’ve been sharing industry news and what you might call tech gossip with you through this newsletter. We decided it’s time to find our most dedicated readers.
On Friday, October 2nd at 7:00 PM, we’re hosting a quiz night at tio-space.
Topic: the last year in the tech world. Products, companies, decisions, crises. The one who knows it all wins. Don’t forget to register here!
Trump Wants to Rename AI
While we were waiting to see whether the US government would grant privileges to major AI companies, restrict AI development, or let them form a cartel, Trump approached the issue from a completely different angle and seems to have decided that the name “artificial intelligence” needs to change first. 🤭
He shared his decision in an X post, saying that AI has gone beyond being “artificial” and that its superiority should be expressed more elegantly, then launched a poll. The names he proposed were: Superior Intelligence, Extreme Intelligence, Supreme Intelligence
According to the poll results, Superior Intelligence appears to be winning with 52%. We’ll see whether Trump succeeds in this rebranding effort while the world burns around him. 😄
Trump’s X Post → | Poll Results →
Jev: The Language Model That Decides Instead of Generates
Diego Almeida, a former OpenAI researcher and one of the creators of ChatGPT, has introduced the first model from his new venture Typesafe AI: Jev. Unlike other language models, Jev doesn’t generate text; it returns a decision from a given set of options, making it a probability model. You can ask it three types of questions: multiple choice, scoring, or true/false. Since the answer can only be one of the provided options, it becomes impossible for the model to hallucinate or produce nonexistent answers. It responds 40 to 200 times faster than other language models while being 400 times cheaper to run.
It’s a particularly compelling solution for agentic AI workflows, where assistants need to choose between tools; these scenarios are nearly deterministic in practice, yet no dedicated alternative existed until now. The concept drew enormous attention: within 24 hours of its announcement, Jev had seen twice the usage of all previous AI models combined. It’s available to try for free via Vercel AI Gateway until September 25th.
Meta Muse: Pick a Side
Amazon blocked Meta‘s AI assistant Muse from shopping on Amazon.com. Shopify, by contrast, went the other direction and integrated Muse into Shop Pay. As AI assistants become more capable of acting on behalf of users, adapting traditional e-commerce, online booking, and similar systems to this new flow becomes increasingly important. Blocking assistants entirely is the easiest guaranteed fix from a security standpoint, but allowing crawls and boosting sales makes more commercial sense. That opens up the question of which crawls to permit, because the same AI assistants that help buyers find the best price can also help sellers scan competitors’ pricing, or detect inventory gaps to raise their own prices.
⚡️ Quick Bits
• Neuralink enabled a patient with a speech disorder to speak in their own voice just by thinking, through an implanted device. Neuralink (X) →
• Since the Hugging Face hack, insiders allege that AI companies have been overstating security breaches to pressure governments into regulation. NY Post →
• Anthropic, whose CEO Dario Amodei called for everyone to slow down just last week, has announced its new model Opus 5.5. Claude AI (X) →
📩 Liked this newsletter? You can forward it to your friends
if you are feeling nostalgic, though it’s always better if
they just subscribe on their own.
