SnackOnAI Blog

Oct 16, 2026

Google's Gemini Argon Turns the Chat API Into a Job Queue

The most important change in Google's new frontier model is not a benchmark. It is that a single answer can now take fifteen minutes, four server side hops, and a polling loop.

Oct 15, 2026

Liquid AI's Agent Model Barely Uses Attention

The fastest way to build a local agent model is not a better transformer. It is a model where attention is almost an afterthought.

Oct 14, 2026

SubQ Small Reads Everything and Attends to Almost Nothing

Long context was never a context window problem. It was an attention budget problem, and Subquadratic's answer is to give every token a fixed one.

Oct 14, 2026

TencentDB Agent Memory Is a Permission System Wearing a Memory Costume

Your agents don't have a recall problem. They have a custody problem: nobody knows who owns what they learned, which version is still true, or which agent is allowed to see it.

Oct 12, 2026

REA: The Agent That Reads Binaries Is Mostly Plumbing, and That Is Why It Works

The AI is the least important part of the most starred reverse engineering tool on GitHub.

Oct 7, 2026

The Framework That Assumes Your Model Is An Adversary

Most agent sandboxes ask whether the model solved the task. Karotte starts from the premise that a model trained against your environment for ten thousand runs will eventually try everything, including the symlink that makes your grader read the answer key to itself.

Oct 5, 2026

The Postgres Predicate That Reads Your Table Twice

A WHERE clause that calls a model once per row would be useless. pg-jev makes it fast by secretly scanning your table ahead of the executor, and that read-ahead window, not your LIMIT, is what you actually pay for.

Sep 30, 2026

Jev Gives Up Writing And Gets A Type System

A model that returns no text cannot hallucinate a citation. It can still hand your code a confidently wrong enum, and the confidence score it ships with is a formula you can write in one line.

Sep 19, 2026

MetaGPT Wrote The Org Chart Into The Prompt, Then Quietly Retired Half Of It

The paper that taught agents to pass documents instead of chatting is still right about the documents. Its own repository no longer believes in the assembly line.

Sep 18, 2026

BMAD Method Compiles Its Prompts, In Five Skills Of Thirty

The flagship skill loads four hundred tokens to reach nineteen thousand, because a Python renderer builds the workflow on demand. Twenty five of the thirty skills still inline everything.

Sep 16, 2026

Orca Supports Twenty Nine Agents With One Terminal

Twenty nine agents needed a terminal. Windows process spawning needed six dedicated documents. The orchestration was never the hard part.

Sep 15, 2026

DeepSeek Harness Ships Hot Reload In One Profile Of Five

Everything is a plugin, including the agent loop. Plugin discovery is a GitHub topic search. One runtime invariant is the only thing standing between those two facts.

Sep 12, 2026

Obscura Runs Every Page Through One JavaScript Thread

Thirty megabytes against two hundred is not a smaller Chrome. It is a different browser that implements less, serializes more, and needed six safety mechanisms to survive the choice.

Sep 11, 2026

llmfit's Fallback Estimate Can Be Fourteen Times Too Slow

Decode is memory bound and prefill is compute bound, so Q4 halves one and leaves the other exactly where it was. Both numbers come from formulas the project publishes, which is what makes them checkable.

Sep 3, 2026

OSWorld's Hardest Baseline To Beat Is An Agent That Gives Up

Thirty of the three hundred sixty nine tasks are impossible. An agent that answers "cannot be done" to every single one scores higher than twenty one of the twenty six baselines in the paper.

Sep 2, 2026

Model Optimizer's Own Benchmark Says Four Bit Is Slower

INT4 AWQ runs Llama3.1-8B seventeen percent slower than the unquantized baseline at batch sixty four, and the number is in NVIDIA's repo.

Sep 1, 2026

CUA-Lite's Parallelism Number Is Just The Memory Ratio

Frontier agents passed OSWorld's human baseline sometime in the last year. CUA-Lite is not built to measure that. It is built to make the training loop cheap.

Aug 30, 2026

OpenSEO Is Very Good At Three Percent Of AI Citations

AI engines cite peer and competitor pages twenty six times more often than they cite your own domain. Every SEO tool you own works on the domain.

Aug 29, 2026

Anthropic's Alignment Agents Did Not Out-Think Anyone

A full automated research run costs about what one human idea costs, and produces a hundred and seventy five of them. That is the whole result.

Aug 26, 2026

Qwen3-VL And The Multimodal Tax

An eight billion parameter model with tools beat the flagship without them. The report has the numbers and does not connect them.

Aug 25, 2026

Rex-Omni, Or What Reinforcement Learning Actually Fixed

A day of GRPO on eight GPUs delivered the headline gain. It improved tight box precision by one tenth of a point.

Aug 24, 2026

FreeToken, Or Why Your Laptop Was Never The Bottleneck

An eight gigabyte laptop GPU just decoded a 35B MoE model at ninety two percent of an RTX 4090's rate, and the reason has nothing to do with the GPU.

Aug 11, 2026

LinkedIn Benchmarked Nine Memory Systems For Its Hiring Agent, Seven Lost To Pasting Everything Into The Prompt

LinkedIn published a table this spring that quietly indicts most of the agent memory market. Their own paper, their own benchmark, their own baselines. Almost nobody has read the rows that matter.

Aug 9, 2026

Autograd Is Slow Per Operation And Free Per FLOP, And That One Sentence Tells You When To Use It

The reverse-mode engine that taught the field how to build autodiff is under two hundred lines of Python. You can read all of it in an afternoon. What you cannot do is make it fast, and the reason is architectural, not a missing optimization.

Aug 8, 2026

The Seedance Paper Is a Scoreboard, Not a Blueprint

ByteDance shipped a twenty six page paper on Seedance 2.0 containing twenty three pages of benchmark tables and zero equations.

Mohinish S

Serverless Ventures | Cloud, Data & Distributed Systems | Angel & Advisor | Infra & Data Startups