🤖 AI Gateway - Lovable AI Gateway Now Free to Try

Lovable AI Gateway is now free to try, and it's a great opportunity for developers to build AI-powered features. From now through Sunday, 9/27 23:59 UTC, it's free to build AI-powered features with Jev. This is a chance to explore the capabilities of Lovable AI Gateway and see what you can ship.

Key Points:

Lovable AI Gateway: Lovable AI Gateway is a platform that allows developers to build AI-powered features. It provides a range of tools and resources to help developers get started with AI development.
Free Trial: The free trial period allows developers to try out the platform and see its capabilities firsthand. This is a great opportunity for developers to explore the platform and determine if it's a good fit for their needs.
Jev: Jev is a model that can be used to build AI-powered features. It's a powerful tool that can help developers create complex AI models.

🔗 Resources:

Original post ↗
Lovable AI Gateway
Jev
Try Lovable AI Gateway for free

🚀 Closed-Loop Provider Failover with Sentry

Closed-loop provider failover is a critical feature for any system that relies on multiple providers. In this post, @fashn_ai explains how they built closed-loop provider failover on top of Sentry's API and webhooks.

Key Points:

Closed-Loop Provider Failover: Closed-loop provider failover is a feature that allows a system to automatically switch to a different provider if the primary provider fails. This is a critical feature for any system that relies on multiple providers.
Sentry's API and Webhooks: Sentry's API and webhooks provide a powerful way to build closed-loop provider failover. By using Sentry's API and webhooks, developers can create a system that automatically switches to a different provider if the primary provider fails.
@fashn_ai's Implementation: @fashn_ai's implementation of closed-loop provider failover is a great example of how to use Sentry's API and webhooks to build this feature.

🔗 Resources:

Original post ↗
Sentry
@fashn_ai
Closed-loop provider failover with Sentry

🚀 4 New Flower Hub Apps

Flower Hub is a platform that allows developers to build and deploy AI models. In this post, @flwrlabs announces 4 new Flower Hub apps that were added during the Collaborative Agent Hackathon in Berlin.

Key Points:

Flower Hub: Flower Hub is a platform that allows developers to build and deploy AI models. It provides a range of tools and resources to help developers get started with AI development.
Collaborative Agent Hackathon: The Collaborative Agent Hackathon was a event where developers came together to build and deploy AI models using Flower Hub.
4 New Apps: The 4 new apps that were added during the hackathon are a great example of the power of Flower Hub. They demonstrate the capabilities of the platform and show how it can be used to build and deploy AI models.

🔗 Resources:

Original post ↗
Flower Hub
@flwrlabs
4 new Flower Hub apps

🤖 AI for Molecular Dynamics

AI for molecular dynamics has a data problem. The trajectories you need to train on are expensive. EGInterpolator (ICLR 2026, with Stanford) learns molecular structure first from abundant conformer data, then uses scarce MD data to learn motion.

Key Points:

EGInterpolator: EGInterpolator is a model that learns molecular structure first from abundant conformer data, then uses scarce MD data to learn motion. This is a powerful tool for AI for molecular dynamics.
Data Problem: The data problem in AI for molecular dynamics is a critical issue. The trajectories you need to train on are expensive, and this can make it difficult to train models.
Stanford: Stanford is a leading research institution in the field of AI for molecular dynamics. Their work on EGInterpolator is a great example of the power of AI for molecular dynamics.

🔗 Resources:

Original post ↗
EGInterpolator
Stanford
AI for molecular dynamics

🤖 Training and Evaluation on Lambda GPU Infrastructure

Training and evaluation ran on Lambda GPU infrastructure. Paper: https://arxiv.org/abs/2604.03911 ↗ v1 … Blog: https://lambda.ai/blog/eginterpo ↗ lator-molecular-dynamics-drug-discovery?utm_source=twitter&utm_medium=organic-social&utm_campaign=2026-09-structures-to-dynamics&utm_content=post-1 …

Key Points:

Lambda GPU Infrastructure: Lambda GPU infrastructure is a powerful tool for training and evaluating AI models. It provides a range of tools and resources to help developers get started with AI development.
EGInterpolator: EGInterpolator is a model that learns molecular structure first from abundant conformer data, then uses scarce MD data to learn motion. This is a powerful tool for AI for molecular dynamics.
Paper and Blog: The paper and blog post provide a detailed explanation of the work on EGInterpolator. They demonstrate the power of the model and show how it can be used to solve complex problems in AI for molecular dynamics.

🔗 Resources:

Original post ↗
Lambda GPU Infrastructure
EGInterpolator
Paper and blog post

🤖 Agents in Production

Agents changed the stack. The stack changed agents. Intelligence without memory is a demo. Memory without a visible path is a guess. We'll walk the system live this Thursday at the Google Chicago office with @GoogleCloud and @MongoDB .

Key Points:

Agents: Agents are a type of AI model that can be used to solve complex problems. They provide a range of tools and resources to help developers get started with AI development.
Stack: The stack refers to the underlying infrastructure that supports the agents. This can include things like databases, APIs, and other tools.
Intelligence without Memory: Intelligence without memory is a demo. This means that the agents are not able to learn or remember anything.
Memory without a Visible Path: Memory without a visible path is a guess. This means that the agents are not able to use their memory to make decisions.

🔗 Resources:

Original post ↗
Agents
Stack
Intelligence without memory
Memory without a visible path

🤖 Proof of Quality

Every evaluation pass in a robotics programme ends in a decision: train on this batch of demonstration episodes, or send part of it back. What a team can say about that decision a month later depends on what the pass wrote down. Each pass in Proof of Quality, Sapien's evaluation system, writes down a record of its decision-making process.

Key Points:

Proof of Quality: Proof of Quality is a system that allows teams to evaluate their AI models. It provides a range of tools and resources to help teams get started with AI development.
Evaluation Pass: The evaluation pass is a critical component of the Proof of Quality system. It allows teams to evaluate their AI models and make decisions about how to improve them.
Decision-Making Process: The decision-making process is a critical component of the Proof of Quality system. It allows teams to make informed decisions about how to improve their AI models.

🔗 Resources:

Original post ↗
Proof of Quality
Evaluation pass
Decision-making process

🤖 Jev Model with Typesafe AI

We were thrilled to get early access to @typesafeai 's Jev model and explore tons of new use cases! Here's what @ktaletsk is building with Jev: 1. jevframe - pandas and polars dataframe adapter to bulk generate new columns with Jev https://github.com/ktaletsk/jevfr ↗ ame …

Key Points:

Jev Model: The Jev model is a powerful tool for building AI models. It provides a range of tools and resources to help developers get started with AI development.
Typesafe AI: Typesafe AI is a leading provider of AI models. Their Jev model is a great example of the power of AI models.
jevframe: jevframe is a tool that allows developers to bulk generate new columns with Jev. This is a powerful tool for building AI models.

🔗 Resources:

Original post ↗
Jev model
Typesafe AI
jevframe

🤖 Adaptive UIs with Jev

Adaptive UIs. Jev really shines in choosing UI elements depending on the context. Interactive marimo-pets widget knows which cell you're viewing and ranks the most useful tools for that code cell. Try here: https://github.com/ktaletsk/marim ↗ o-pets/tree/experimental/ambient-palette …

Key Points:

Adaptive UIs: Adaptive UIs are a type of user interface that can be customized based on the user's context. This is a powerful tool for building user-friendly interfaces.
Jev: Jev is a model that can be used to build adaptive UIs. It provides a range of tools and resources to help developers get started with AI development.
Interactive Marimo-Pets Widget: The interactive marimo-pets widget is a great example of the power of adaptive UIs. It knows which cell you're viewing and ranks the most useful tools for that code cell.

🔗 Resources:

Original post ↗
Adaptive UIs
Jev
Interactive marimo-pets widget

🤖 Landing First Customers

Halfway through my interview with a YC founder, I had to stop him. I was sure I'd misheard. I'd asked how he landed his very first customers. His company had hit $3.2M ARR in 11 months. His answer: "We used 7 slides to present our idea. At the end, we asked people to pay, even though we didn't have a product yet."

Key Points:

Landing First Customers: Landing first customers is a critical step for any startup. It requires a clear and compelling pitch that can be used to convince potential customers to try the product.
YC Founder: The YC founder's answer to the question of how he landed his first customers is a great example of the power of a clear and compelling pitch. He used 7 slides to present his idea and asked people to pay, even though he didn't have a product yet.
$3.2M ARR: The company's $3.2M ARR in 11 months is a testament to the power of a clear and compelling pitch. It shows that it's possible to land first customers and build a successful business.

🔗 Resources:

Original post ↗
Landing first customers
YC founder
$3.2M ARR