Stop Blaming Serialization for Slow Python UDFs
The benchmark that tested the wrong Arrow Earlier this year I built a benchmark to answer a question that comes up constantly: if you need custom Py…
Latest News
⚑ Report a ProblemTech news from the best sources
All topics
AI
Gear
News
Tech
agents
ai
api
architecture
automation
beginners
career
database
devchallenge
devops
discuss
javascript
llm
machinelearning
mcp
opensource
performance
productivity
programming
python
security
showdev
testing
tutorial
typescript
webdev
The benchmark score is the number to trust least
Claude Opus 5.5 scores 58 on Artificial Analysis's Intelligence Index, more than double the median of 26 for comparable reasoning models. That is th…
The Visible Test Suite Is the Ceiling on What the Next Agent Can Deliver
Originally published at tddbuddy.com . Related reading: Coverage Was a Proxy. The Proxy Just Failed. names mutation score as one metric under the ce…
Wild beats mold linking Rust 20 times out of 20, and in release the linker is no longer the bottleneck
I measured Wild 0.10.0, mold 2.42.1, rust-lld and GNU ld linking ripgrep and cargo. Wild won all 20 repetitions of each project, by 1.9 ms and about…
TabPFN and TabICL against tuned XGBoost: the model that does not train won on fourteen tables out of fourteen
The claim behind TabPFN and TabICL: they predict on a table without ever training on it, and still beat tuned boosting. Measured on 14 datasets from…
Rust Coreutils 0.12 vs GNU 9.12: almost everything works the same, starting a process costs nearly 3x more
I measured Rust Coreutils (uutils) 0.12 against GNU coreutils 9.12 in containers with limits: the full GNU test suite, 157 commands with their outpu…
Efficiency Hallucination: Every Model Rewrote Code That Couldn't Get Faster
Okay, this is going to sound dumb, but I spent Sunday evening asking three different models to make a function faster that could not be made faster,…
Giving a coding agent more time barely helps
Real-SWE ran frontier models against licensed, private enterprise codebases (billing, tax, multi-service work) and one number jumped out at me: roll…
GPT-6 Astra scores 95% on one robot task, 10% on another
Robocurve has run OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1 against physical robot arms, and the results split sharply by task. Astra co…
An AI Capture-the-Flag Tournament: What the Scoreboard Counted
Code: Megapixel99/capture-the-flag In April I ran five games of an AI capture-the-flag tournament between five small open-weight models (1.0B to 2.5…
We hit 99.95% on the LoCoMo memory benchmark. Here's the catch, and why it still matters.
Our CEO Rob Imbeault published a piece on LinkedIn this week about a result our team posted: 99.95% on LoCoMo , the most cited benchmark for long-te…
How I tried to write an article about slow Chinese LLMs
Recently, I've added a bunch of hype-monsters to my AI Werewolf : Kimi K3 Qwen 3.8 Max, Qwen 3.7 Plus, Qwen 3.7 Flash MiniMax M3 Plus the ones I've…
AI News Roundup: Grok 4.5 Hits Tesla, Perplexity's Orchestrator Beats Opus, and Meta Undercuts Pricing
Five stories moved the AI-coding world today. None are about a single model winning forever — they are about the ground shifting under who runs the…
Turn the camera away, and the AI's world freezes
Video AI systems consistently fail to track what happens when the camera looks away: when a scene pans away from an object in motion and returns, cu…
Reliable, and still wrong
A large-scale audit of AI-as-judge evaluation — covering over half a million individual judgments — finds that AI judges are consistently reliable b…
Claude Fable 5 Scores 95% on SWE-bench, Then Hands Off to Opus 4.8
The headline number is 95% on SWE-bench Verified. That's the score attached to Claude Fable 5, Anthropic's new general-access model in the Mythos cl…
Cross-Machine Memory Query: About 20 Milliseconds, Most Days
I wrote about hardware benchmarks twice this week. Different problem this time. Same machines. I have a Mac for daily work, a Linux box that runs a…
An AMD GPU Beat My Mac on Llama 8B. The Same GPU Lost on Phi-3.
I wrote a post yesterday about why GPUs barely help small text embeddings at batch=1. Different workload, same machines. This time I ran a local LLM…
Your GPU Probably Isn't Helping Your Retrieval System
Most "just use a GPU" advice is wrong for how anyone actually runs small models. I spent yesterday benchmarking a 33M parameter embedding model acro…
pypdf vs PdfPig: Text Extraction at Scale
Overview PDF text extraction is a common pre-processing step in data pipelines — ingesting research papers, legal documents, or reports before embed…
NetworkX vs CSR + TensorPrimitives: PageRank on 28M Edges
Overview PageRank is the canonical graph algorithm. NetworkX implements it in pure Python — its dict-of-dict adjacency representation means every po…
SurrealDB 3.x by the numbers
Author: Tobie Morgan Hitchcock One engine, multi-workloads, full durability. You can explore the full results, methodology, and per-database breakdo…
What ground truth caught that unit tests missed: 3 real bugs in 9 flagship lint rules
We added a npm run ilb:flagship:smoke gate to the quality script. It's small: for each flagship rule with a labeled corpus, run the rule against vul…
When Generic Benchmarks Fail: Building a Sales-Domain Evaluation Bench from Scratch
By Natnael Alemseged The gap that τ²-Bench retail cannot measure Tenacious is a B2B sales automation company. Its agent produces outreach emails for…