What people are shipping with Jev.

Ayer en stream nos pusimos a programar un bot que juega Slay the Spire 2 automaticamente usando Jev de @typesafeai Por ahora solo tiene la parte de combate, pero funciona bastant

@KyleHovance @typesafeai Real-time agents: LLM sets high-level goals while Jev evaluates state every ~100ms to pick typed actions (move, shoot, dodge) without text gen delays. Sup

@obie @typesafeai Apart from speed and cost, are there any benchmarks that compare jev to LLMs on known tasks that we care about?

apparently the most useful object is a smartphone 📱 @typesafeai Jev ranked 1,000 everyday objects it beat refrigerator 51–49 2.8s. <1¢. #AI https://t.co/nBUO3CDCCT

@juanseb @tamarajtran @typesafeai El post de tamarajtran propone usar Jev de typesafeai para compactación instantánea de contexto: puntúa cada tool call y descarta irrelevantes en

JEV Plays Runescape Set up Jev on an 2004 OSRS Server running 5 bots at once. The goal? Highest total level possible. Will be optimising this to let them start PVP'ing soon.

@pedronauck @typesafeai Nice, this is the other half of what I've been building. You use Jev to decide what leaves the context, I use it to decide what enters, jev-rules asks Jev w

Antes de usar Jev de @typesafeai con datos de tu empresa, lee sus términos: - "No envíes información confidencial" - Responsabilidad máxima: 100 $ - Sin DPA, sin SOC 2, sin lista d

We just shipped @typesafeai's Jev to prod as our AI Assistant's reranker. (it's under feature flag only for selected customers still) The early numbers are promising: ⚡ 3.3× faste

We are finally starting to see the feedback loop play out. The reaction to AI-slop machines. The tools are being built. Now we have programming languages designed to deal with hal

Jev - realtime youtube comments classification through a chrome extension. @typesafeai https://t.co/oDZuiaWV8G

Big thanks to @TypeSafeAI for access to Jev. It’s pushing me to rethink AI systems: deterministic code first, typed decisions where uncertainty exists, and LLMs only when deeper r

@spacepixel @typesafeai yes been testing it and so much that can be done via browser automations and polling, JEV is so damn fast

@typesafeai Before shipping we replayed real user questions through it: Jev put the right docs higher (nDCG@5 0.65 vs 0.56, graded by an LLM judge).

#FF Jev @typesafeai

Fun to use @typesafeai ‘s Jev model to make an efficient decision on what cloud region to failover to in an incident using the @gocutover task graph (controls, deterministic autom

If you're building something with Jev by @typesafeai don't ask it one broad question. Break it down into narrower questions and then tally the results. Here's a practical example.

Busca em base de dados de 20k itens pra encontrar itens que batem com o que o usuário pediu O jev da @typesafeai vai me economizar uma boa grana e tempo do usuário https://t.co/Gd

Yes, We integrated the @typesafeai Jev in the SuperQode with Unique use case. Jev in SuperQode: a harness for your code and workflows with Jev agent 📙 Read: https://t.co/WY0uZR

Built a @typesafeai plugin for @openclaw. It adds one optional OpenClaw tool for typed Jev decisions: supply state plus Noul, Choice or Score questions and get answers from a defin