AI in the AM โ€” Weekly Highlights: Relaunch Week (Aug 17โ€“20, 2026)

Cognitive Revolution "How AI Changes Everything" ยท 3K views

AI for Emergency Response + OpenAI's Real-Time Voice Lead

Cognitive Revolution "How AI Changes Everything" ยท 162 views

Gemini 3.7 Flash

Meta's new model wants "deep access" to your personal life...

Fireship ยท 485K views

I spent 3 days at MIT... the robot hype is worse than you think

Fireship ยท 1.1M views

Lindy Teammate: Flo Crivello on Multiplayer Agents, Memory & Why He'd Ban the Chinese Models He Uses

Cognitive Revolution "How AI Changes Everything" ยท 926K views

Thinking in Silico: Goodfire CTO Dan Balsam on Concept Manifolds & a $1000/Month ML Research Agent

Cognitive Revolution "How AI Changes Everything" ยท 1.2M views

Pick Your Poison: Zvi Mowshowitz on the Unipolar/Multipolar AGI Dilemma, OpenFace & Pacing the ...

Cognitive Revolution "How AI Changes Everything" ยท 3K views

Nathan Goes to China โ€“ Part 2: AI Safety with Chinese Characteristics

Cognitive Revolution "How AI Changes Everything" ยท 1K views

Claude Opus 5

claude
Context
1M
Max out
128K
In $/M
$5
Out $/M
$25
API ID
claude-opus-5
Mid-conversation tool changes are now in beta on several Claude models.2026-07-24
Available in beta on Claude Fable 5, Claude Mythos 5, Claude Opus 4.8, and Claude Opus 5, allowing adding or removing tools between turns while preserving the prompt cache via the mid-conversation-tool-changes-2026-07-01 beta header.
docs.claude.com โ†—
Effort is now the primary control for steering Claude Opus 5.2026-07-24
The model supports the full effort ladder (low, medium, high, xhigh, max), with max for capability-critical work.
docs.claude.com โ†—
On Claude Opus 5, disabling thinking is allowed only at effort high or below.2026-07-24
Using thinking disabled with effort xhigh or max returns a 400 error, a breaking change from Claude Opus 4.8.
docs.claude.com โ†—
Dreams now supports Claude Opus 5.2026-08-01
Dreams (research preview) now supports Claude Opus 5. Details are in the Supported models documentation.
docs.claude.com โ†—
Claude Opus 5 was launched as a step-change improvement over Claude Opus 4.8.2026-07-24
Supports a 1M token context window (both default and maximum), 128k max output tokens, and thinking on by default, priced at $5/$25 USD per MTok, the same as Claude Opus 4.8. Available on the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.
docs.claude.com โ†—
โ— No recent changes

Gemini Robotics ER 2 (streaming)

Gemini Robotics ER 2

Is Offense or Defense Dominant? FAR.AI's Adam Gleave on the AI Security Leaderboard

Cognitive Revolution "How AI Changes Everything" ยท 63K views

Did Anthropic just kill the indie hacker...?

Fireship ยท 699K views

GPT Live Transcribe

openai

GPT Transcribe

openai

Nathan Goes to China โ€“ Part 1: Tech & Agent Setup, Chinese AI UX, WAIC, and Attitudes on AI

Cognitive Revolution "How AI Changes Everything" ยท 65K views

Gemini 3.5 Flash-Lite

Gemini 3.6 Flash

Codex CLI

Terminal coding agent ยท openai
Vendor
OpenAI
Surface
CLI
Open source
Yes

OpenAI's open-source, terminal-native coding agent. Codex CLI runs locally, reads and edits your code, and executes commands in a sandbox to complete multi-step tasks. Pairs with OpenAI models.

โ— No recent changes

OpenClaw

Self-hosted AI assistant gateway ยท openclaw
Vendor
OpenClaw
Surface
Self-hosted gateway
Open source
Yes

An open-source, self-hosted personal AI assistant. OpenClaw is a gateway that connects chat apps (WhatsApp, Telegram, Discord, Slack, and more) to AI agents, so you can drive an agent โ€” including coding and GitHub workflows โ€” from a messaging app on any platform.

โ— No recent changes

Claude Code

Coding agent ยท CLI + desktop + IDE ยท claude

opencode

Terminal coding agent ยท sst

Cline

Autonomous coding agent ยท VS Code ยท cline
Vendor
Cline
Surface
VS Code
Open source
Yes
โ— No recent changes

Claude Fable 5

Frontier reasoning ยท long-horizon agents ยท claude

Hermes Agent

Personal AI agent ยท CLI + gateway ยท nous

Goose

On-machine AI agent ยท CLI + desktop ยท block

Continue

AI code assistant ยท IDE ยท continue
Vendor
Continue
Surface
IDE extension
Open source
Yes

Open-source AI code assistant for VS Code and JetBrains โ€” inline autocomplete, chat, and edit features wired to the models and context sources you choose. Built as configurable building blocks rather than a fixed product.

โ— No recent changes

Aider

AI pair programming ยท terminal + git ยท aider
Vendor
Aider
Surface
CLI
Open source
Yes

Open-source AI pair programming in the terminal. Aider edits code across your local git repository, makes coordinated multi-file changes, and auto-commits each edit with a sensible message. Model-agnostic โ€” works with Claude and other frontier models.

โ— No recent changes

Claude Haiku 4.5

Fastest ยท near-frontier ยท claude
Context
200K
Max out
64K
In $/M
$1
Out $/M
$5
API ID
claude-haiku-4-5

Anthropic's fastest model with near-frontier intelligence โ€” the low-latency, low-cost pick for high-volume work. 200K-token context with 64K max output; supports extended thinking. Priced at $1/$5 per MTok.

โ— No recent changes

Claude Desktop

Desktop app for Claude ยท claude
Vendor
Anthropic
Surface
Desktop app
Open source
No

Anthropic's desktop application for Claude (macOS and Windows). Beyond chat, it supports Model Context Protocol (MCP) servers, so Claude can connect to local tools, files, and external services from the desktop. Proprietary.

โ— No recent changes

Cursor

AI code editor ยท cursor
Vendor
Anysphere
Surface
Editor
Open source
No

A popular AI-first code editor (a VS Code fork) from Anysphere. Cursor builds agentic editing, multi-file changes, codebase-aware chat, and predictive autocomplete directly into the editor. Proprietary; works with frontier models including Claude.

โ— No recent changes

Windsurf

Agentic AI IDE ยท windsurf
Vendor
Windsurf
Surface
Editor
Open source
No

An agentic AI IDE built around 'Cascade,' an agent that reads, edits, and runs across your codebase with awareness of your recent actions. An AI-native editor in the same category as Cursor. Proprietary.

โ— No recent changes

Claude Opus 4.8

Frontier reasoning ยท agentic coding ยท claude
Context
1M
Max out
128K
In $/M
$5
Out $/M
$25
API ID
claude-opus-4-8

Anthropic's most capable Opus-tier model, built for complex reasoning, long-horizon agentic coding, and high-autonomy work. 1M-token context with 128K max output; adaptive thinking is on and the effort parameter defaults to high on all surfaces (including Claude Code). Priced at $5/$25 per MTok.

โ— No recent changes

Claude Sonnet 4.6

Balanced speed + intelligence ยท claude
Context
1M
Max out
64K
In $/M
$3
Out $/M
$15
API ID
claude-sonnet-4-6

Anthropic's best combination of speed and intelligence โ€” the everyday workhorse for production workloads. 1M-token context with 64K max output; supports both extended and adaptive thinking. Priced at $3/$15 per MTok.

โ— No recent changes