Reducto.ai is adding a pricing calculator. The calculator shows potential costs as document pipelines scale. Users can test it with their own numbers at the provided URL.
Key points
Feature: Pricing calculator added to estimate pipeline costs.
Access: Try the calculator at https://reducto.ai/pricing.
Sources
- Original post
- Linked resource - Linked in the post
Pally AI launches photo‑based macro tracking (Day 9)
Pally AI announced a new feature on Day 9 of its launch week. Users can take a photo of meals and Pally will track calories and macros. The feature has been popular in recent months and has been optimized.
Key points
Launch week Day 9: macro tracking via photo.
Optimized for tracking calories and macros from meal photos.
Sources
AI Popup Retrigger shows 23% increase in opt‑ins
AI Popup Retrigger is one of five AI products a 9‑figure brand uses on PostscriptIO. A holdout test showed a 23% increase in opt‑ins for that product.
Key points
AI Popup Retrigger: 23% increase in opt‑ins
Brand uses 5 AI products on PostscriptIO
Sources
Pally AI adds photo‑based macro tracking
Pally AI lets users photograph meals to log calories and macros. The feature supports fitness goal tracking. It has been a popular use case in recent months and has been optimized.
Key points
Feature: Photo capture of meals for automatic calorie and macro logging.
Update: Optimized for the popular macro‑tracking use case.
Sources
Top 100 Consumer AI Apps report released; SierraPlatform and DecagonAI launch
a16z and omooretweets released the Top 100 Consumer AI Apps report. SierraPlatform and DecagonAI launched their own offering. Both announcements were made within the past four days.
Key points
Report: a16z & omooretweets released the Top 100 Consumer AI Apps report
Launch: SierraPlatform & DecagonAI launched their own
Sources
Datapoint Streams – online RL with live human feedback for multimodal models
Datapoint Streams provides online reinforcement learning that incorporates real‑time human preferences for image, audio, and video models. The system sends rollouts to a pool of over one billion raters, receives structured preferences as a continuous reward stream, and updates the policy while training. Low‑trust responses are filtered before influencing updates.
Key points
1B+ raters judge rollouts in real time, delivering preferences as a continuous reward stream.
Feedback latency is measured in seconds per judgment, with low‑trust answers dropped in flight.
Sources
- Original post
- Streams: online RL with live human feedback - Linked page
LVMH partners with Analog Devices on AI grape spoilage sensor
LVMH has partnered with semiconductor software firm Analog Devices to create a device that uses AI to detect the aroma of spoiled grapes. The device is trained on previous bad crops from Moët Hennessy wine. It builds a chemical “fingerprint” of the smell, referred to as Fresh Mushroom Aroma.
Key points
Partnership: LVMH and Analog Devices collaborate on the device.
Training data: Device trained on bad crops from Moët Hennessy wine.
Fingerprint: Chemical “fingerprint” named Fresh Mushroom Aroma.
