Culture · Products · Signal

Salt Index

There are more products, platforms and services than ever before.

Read the indexAbout
Sections
apisartificial-intelligencecryptocurrencyfinanceinvestingsoftwaretechnologyweb-scraping
Gemini Omni 1.2 Flash GA Brings Longer Clips and a Steeper Price Cut for Video Teams
artificial-intelligence

Gemini Omni 1.2 Flash GA Brings Longer Clips and a Steeper Price Cut for Video Teams

Google DeepMind pushed Gemini Omni 1.2 Flash to general availability with 15-second generations and roughly thirty percent lower per-second pricing at 720p.

A Buyer Guide to Gemini Omni 1.2 Flash for Production Video Workloads
technology

A Buyer Guide to Gemini Omni 1.2 Flash for Production Video Workloads

Google's November 5 GA brings 15-second clips at $0.07 per 720p second, conversational edits, and SynthID watermarks for ads and product video at scale.

OpenAI Images 2.6 Promises 50% Faster Generation and Prompt-to-GIF Workflows
technology

OpenAI Images 2.6 Promises 50% Faster Generation and Prompt-to-GIF Workflows

The refresh lands in ChatGPT, Codex, and the API—here is what changed and who should care.

AMD and Super Micro After a Record AI Infrastructure Quarter: A Buyer and Builder Guide
technology

AMD and Super Micro After a Record AI Infrastructure Quarter: A Buyer and Builder Guide

AMD's data-center surge and Super Micro's raised outlook give teams a snapshot of who wins when AI servers become the default enterprise purchase.

technology

Marvell to Acquire Lumentum for $115B in AI Connectivity Optics Megadeal

Marvell announced a $115B all-stock Lumentum deal Nov 2 to combine electrical and optical AI datacenter connectivity; buyers should dual-source transceivers and lock SLAs through a long close.

artificial-intelligence

Gemini 4 Argon, Carbon, and GPT-6.1 Terra: A Frontier Model Buyer Guide

Which November 2026 frontier tier fits general product work, cyber defense, or cheap computer-use automation.

Qualcomm After the Chip Rout: Apple Modem Risk, Arm Win, and the AI Repositioning Skeptics
technology

Qualcomm After the Chip Rout: Apple Modem Risk, Arm Win, and the AI Repositioning Skeptics

Salt Index buyer guide as QCOM drops with Intel—Huawei licensing, Oryon CPUs, and November 4 earnings keys.

The Index

Deepgram buys Daily and Pipecat, betting on an open voice-agent stack
technology

Deepgram buys Daily and Pipecat, betting on an open voice-agent stack

technology

x402 AI Agent Payments Without Chargebacks: A Buyer Guide

Apple iPad mini With A20 Pro and OLED: 2026 Buyer Review for Travel and On-Device AI
technology

Apple iPad mini With A20 Pro and OLED: 2026 Buyer Review for Travel and On-Device AI

apis

OpenAI Decisions API Hits GA at $0.10 per Million Input Tokens

artificial-intelligence

Kimi K3.1 Pro vs MiniMax M3.1 Flash vs Mercury 3: Which API Fits Your Stack?

Muse Spark Open Self-Hosting Guide: NVFP4, APIs, and Real Hardware Costs
technology

Muse Spark Open Self-Hosting Guide: NVFP4, APIs, and Real Hardware Costs

View all latest posts

Briefing

World ID for Banking Targets $40 Billion AI Fraud With Proof-of-Human Checks
artificial-intelligence

World ID for Banking Targets $40 Billion AI Fraud With Proof-of-Human Checks

World launched a finance-focused World ID suite at Money20/20, bundling app, Zoom, and agent verification for banks fighting deepfakes and bots.

BlackRock Files PBIT, a Bitcoin Managed Protection ETF Built on IBIT Options
investing

BlackRock Files PBIT, a Bitcoin Managed Protection ETF Built on IBIT Options

BlackRock's Oct 19 Form 485APOS outlines PBIT, an IBIT-options wrapper targeting roughly 10% downside buffers via call spreads, with a 75-day path to a January 2027 launch.

Apple Intelligence China Beta Clears Path With Alibaba Qwen and Tongyi Wanxiang
technology

Apple Intelligence China Beta Clears Path With Alibaba Qwen and Tongyi Wanxiang

Late-October trials bring translation and writing tools without AI Siri, as MIIT approvals pair Apple's hardware with locally registered generative models.

Reflection Beam Ships as a 501B-Parameter Apache 2.0 Coding Model With a 1M-Token Context
artificial-intelligence

Reflection Beam Ships as a 501B-Parameter Apache 2.0 Coding Model With a 1M-Token Context

Reflection's first open-weight MoE targets agentic software work with SWE-bench scores and day-zero Google Cloud support.

Across the Network