Skip to content
MustHave.ai
MustHave Hoodie Seal
THE MUSTHAVE NEWSROOM

AI news

What’s new. What matters. What do you think?

Top stories

By member votes

Audible’s TTS Research Targets Consistent Character Voices

Audible describes controllable TTS and book-level character profiles. Review the company’s reported results and how to check voices across distant chapters.

OmniDR-Bench Tests AI Video Reports for Evidence and Citations

OmniDR-Bench evaluates AI reports built around a source video. Review its evidence checks, citation limits and the practical work of auditing a report.

AI21’s LIFT-Bench Tests Whether AI Drafts Follow the Brief

AI21’s LIFT-Bench checks section-level content and formatting. Its October 25 explainer shows how a high reward can hide an incomplete AI draft.

Tencent Releases Hy-SimulMT for Live Chinese-English Translation

Tencent releases Hy-SimulMT for live Chinese-English translation. Review model variants, separate speech output and the documented deployment limits.

What Counts as a Zapier Task? Estimate an AI Workflow’s Usage

Estimate Zapier task usage for basic steps, AI model tiers and MCP calls. Use an explicit workflow example before comparing plan allowances or monetary costs.

How to Send PDF Invoice Tables to Google Sheets With Parseur

Send PDF invoice data to Google Sheets with Parseur. Map line items, check totals and prevent duplicate imports using a clearly defined sample invoice.

Ahrefs Free: What You Can Use Without a Paid Subscription

Use Ahrefs Free for verified-site research and audits. Compare its account access with standalone tools before deciding whether a paid subscription is necessary.

How to Export Instagram Comments to CSV With EasyComment

Export public Instagram comments to CSV or Excel with EasyComment. Check credits, replies and missing rows before using the file for a campaign report.

Kling 4.0 Arrives in Luma: Evaluate One Difficult Camera Move

Luma adds Kling 4.0 to its creative workspace. Test one difficult tracking shot for continuity and usable output before comparing models or production costs.

NVIDIA Minerva: Separate Telemetry, Diagnosis and Automated Repairs

NVIDIA introduces Minerva for AI infrastructure operations. Evaluate telemetry, diagnosis and permission to automate a repair separately.

ThinkingBox Training: Measure Repeated Success After Reinforcement Learning

ThinkingBox releases a training integration based on final database state. Reported gains in repeated success come with a decline in at-least-once coverage.

Unity MCP 1.5 Adds a CLI: Review What the Agent Can Change

Unity MCP 1.5 adds a standalone CLI and skills support. Review read-only resources, editor action tools and authentication before connecting a valuable project.