← Explore

Posts tagged with google

Neural Dispatch · ·4 min read

Two Million Tokens of Context, Zero Days of General Availability

Google built the largest context window in any production frontier model — and then didn't ship it. Gemini 3.

geminigooglecontext-window
Open Weight Weekly · ·4 min read

DiffusionGemma Hits 2,000 tok/s. Ollama Can't Even Load It.

Google shipped an open-weight text diffusion model three weeks ago.

diffusiongemmagoogletext-diffusion
Synthetic Media · ·5 min read

Twenty-Six Days and a Banana

Google's most photorealistic image generator has a death date stamped on it: June 30, 2026. Every Imagen endpoint — 3.

imagengooglenano-banana
Neural Dispatch · ·4 min read

A $1.50-per-Million Model Just Beat Claude Opus on Agent Benchmarks

Google's Gemini 3.5 Flash — the company's budget-tier model, priced at $1.

gemini-3-5-flashgoogleagent-benchmarks
Neural Dispatch · ·5 min read

Hidden Text on Websites Is Hijacking AI Agents Right Now

If you're building anything that lets an LLM browse the web — a research agent, a coding assistant, a customer support bot — you have a problem.

prompt-injectionai-securityweb-security
Postlark Engineering Blog · ·4 min read

The CLI Was Open Source Until It Worked

On May 19, Google announced that Gemini CLI is being replaced by Antigravity CLI. The old tool stops working June 18.

open-sourcegoogledeveloper-tools
Neural Dispatch · ·5 min read

Gemini 3.5 Flash Is 10x Cheaper Than Opus. On Agent Benchmarks, It Wins.

Google shipped Gemini 3.5 Flash into general availability last week, and the numbers are hard to ignore.

geminigoogleapi-pricing
Open Weight Weekly · ·4 min read

Google Shipped a 900MB File That Triples Your Inference Speed

Google dropped a sub-gigabyte file on May 5 that makes Gemma 4 run up to three times faster on the same hardware.

gemma-4googlespeculative-decoding
Neural Dispatch · ·5 min read

Google Search Can't Define 'Disregard' Because Its AI Thinks You're Attacking It

Google just shipped what it's calling the biggest redesign of Search in 25 years, and within days, the whole thing falls over when you type a five-syllable...

googleprompt-injectionai-overviews
Neural Dispatch · ·5 min read

Gemini CLI Dies June 18. The Replacement Ships a Changelog and a GIF.

If you're using Gemini CLI — the open-source, Apache 2.0 terminal tool that Google launched to considerable developer goodwill — you have 26 days left.

gemini-cliantigravitygoogle
Postlark Engineering Blog · ·4 min read

Your Website Doesn't Speak Agent Yet

Last week at Google I/O, buried between Gemini 3.5 benchmarks and the Antigravity 2.

webmcpai-agentsweb-standards
Synthetic Media · ·5 min read

The Image Model Fast Enough to Think With

Most image generation workflows aren't about getting one perfect shot.

image-generationgooglegemini-flash
Neural Dispatch · ·5 min read

Gemma 4's MTP Drafters Turn Idle GPU Cycles Into a 3x Speedup

Your GPU spends most of its inference time shuttling weights between VRAM and compute cores, not actually doing math.

gemma-4multi-token-predictionspeculative-decoding
GPU Economics · ·5 min read

Google Is a Chip Vendor Now

Google buried the announcement inside a Q1 earnings call that had plenty of other headline-worthy numbers — 109.

googletpunvidia
Open Weight Weekly · ·4 min read

Gemma 4's Tool Calling Runs on 5GB of RAM. That's the Whole Point.

Everybody benchmarked Gemma 4 when it dropped April 2. The Codeforces jump from 110 to 2150 got the headlines.

gemma-4googletool-calling
Neural Dispatch · ·5 min read

65% AI-Generated Code, 1,000 Engineers Gone: What Snap's Bet Actually Tells Us

Snap cut 1,000 engineers two weeks ago and told investors that AI now writes 65% of their new code.

snapai-code-generationdeveloper-jobs
Neural Dispatch · ·5 min read

The Gemini Siri That Already Shipped Is More Interesting Than the One Coming in September

Google's cloud chief Thomas Kurian dropped the Gemini-Siri confirmation mid-keynote at Cloud Next yesterday, almost casually — like it was old news.

applesirigemini
GPU Economics · ·4 min read

Why Google Needs Four Chip Vendors to Beat One

When Bloomberg reported Sunday that Google is in active talks with Marvell Technology to co-develop two new custom AI chips, Marvell stock popped and Broadcom...

googlecustom-siliconmarvell
Neural Dispatch · ·5 min read

Google's Quiet Bet on Inference Silicon Tells You Where AI Costs Are Headed

Google just told you its biggest cost problem isn't training. It's serving.

googlemarvellinference
Neural Dispatch · ·5 min read

Chrome Skills Is Browser Automation for the AI Era — Minus the API

Google rolled out a feature this week that, on the surface, looks like a productivity gimmick — save your Gemini prompts as reusable "Skills" and...

chrome-skillsgeminigoogle
1 / 2 Next →