Veille techno
Toute l'actualité tech, sur une page.
L'Essentiel de la Tech
A PhD student in ML and ecology on why faster code, experiments, and drafts don't fix data with a backstory, and three predictions for how labs will change. The post What AI Agents Speed Up in PhD Research, and What They Can't appeared…
Looking for a free alternative to Jev? Here are 5 open-source decision models you can run locally, from high-accuracy options to lightweight models for AI agents.
Gemma 4 E2B drafts the replies, Bespoke Labs' Nimble 9B picks accept, revise, or human review, and only flagged questions get a focused retry. The post How to Check Local LLM Output with Ollama, Gemma 4, and Nimble appeared first on…
This post was created with the assistance of AI, based on a transcript of the conversation. AI tools have started to remember their users, but each tool remembers a different version of the same person. Duncan Davidson, an independent…
A reproducible test of BM25, embeddings, and parent documents shows why author-written evals overrate search, and why answer position in context matters. The post RAG Retrieval Lost 40 Points When Distracting Passages Grew to 2.68M…
Across 2,312 planted faults the study also surfaced four real library problems, from a Windows data-loss bug to README examples that do not work as written. The post I Mutation-Tested 50 Python Libraries. Tests in the Young Ones Missed 41%…
Learn how graph-enhanced RAG connects related facts that plain vector search misses, then build a multi-hop research assistant with AWS Labs' open-source graphrag-toolkit, Neo4j, and pgvector.
Annotators marked each MNIST and Mukhoti digit yes, no, or unsure, producing image-level soft labels and uncertainty scores. The post A Soft-Label MNIST Testbed for Human vs. Model Confidence appeared first on Towards Data Science .
Building an AI agent is easy. Building one that actually works in production is not. These 7 free books cover the engineering knowledge you need to make that jump.
A simulated scar and data from 31 animals after ischemia show how an inhomogeneous estimator, random labelling, and brms separate tissue loss from real interaction. The post How to Read Cross Pair Correlation in R Without Fake Repulsion…
The shift to agentic reality When Git was designed in 2005, Linus Torvalds solved a brilliant problem for humans collaborating over terminal windows by trading line-by-line diffs. Fast-forward to 2026: Autonomous coding agents don’t think…
Abacus.AI applied the same fine-tuning recipe to three very different open-weight models. The gains landed in completely different places & that tells you more than any single benchmark score.
Three small open models, 150 MMLU questions each, five pushback styles: “That’s wrong” did more damage, and naming a wrong option was close to fatal. The post “Are You Sure?” Flipped Qwen2.5-1.5B’s Right Answer 2 Times in 3 appeared first…
Most of today’s AI entrepreneurs and social media enthusiasts are young. AI is the first technology revolution they’ve lived through. They don’t remember how unbelievably exciting it was in the late 1970s to get your own computer, instead…
A palace-hotel wedding week shows where assistants, copilots, agents, and orchestration fit, and why Gartner expects 40%+ of agentic AI projects to be cancelled. The post AI Assistant vs. Copilot vs. Agent, and When to Use Each appeared…
A data-backed breakdown of the skills, credentials, and capabilities that hiring managers are screening for in senior roles right now.
A Sate Padang vision test and a Malaysia capital question show why the free, private model needs grounding and validation before it leaves prototyping. The post How to Use Apple's fm CLI to Run On-Device LLMs From Python appeared first on…
Learn how to measure LLM inference performance using key metrics and formulas for latency, token generation speed, and system load.
The following article was originally published on James Shore’s Successful Software blog and is being republished here with the author’s permission. AI is great for making teams faster. But is it sustainable? What about maintainability…
Trained on six PathMNIST tissues with a small KEEP adapter, it handled added labels without retraining, though a linear classifier hit 88% on familiar ones. The post A Jev of All Trades: Build and Train a Jev-style Decision Model From…
Learn how AI can make enterprise business intelligence more accessible, automate reporting, uncover insights faster, and improve decision-making with the right governance.
Jev needs no training and at most three examples per intent; the open rebuild needs labels. Both still fall into the none-of-the-above trap. The post Scikit-Learn Matched or Beat Jev's Calibration on BANKING77 appeared first on Towards…
A software tester ran six testing skills on Haiku, Sonnet, and Opus; pushy descriptions added only 5 points, and near-miss requests never fired a skill. The post A Session-Start Hook Raised Claude Code Skill Firing From 15% to 62% appeared…
Learn, in a few easy steps, how to build a two-stage scikit-learn pipeline for multilingual text classification, guided by two pre-trained language models handled by Scikit-LLM for translation and zero-shot classification.
A Gemini Flash-Lite step pulls skills, seniority, and Nigeria eligibility from 18 RemoteOK postings, with Python checks and dbt tests guarding every answer. The post How to Validate LLM Output in a DuckDB and dbt Pipeline appeared first on…
Learn five practical ways to fine-tune LLMs on a budget, from LoRA and QLoRA to Unsloth, DPO, and right-sizing your GPU, with working code for each technique.
The following article originally appeared on the Asimov’s Addendum Substack and is being republished here with the author’s permission. This essay is based on my conversation with Demetrios Brinkmann for AAIF Live, recorded before my…
A guard on negative enrollment days looked correct while the function it called overbilled a customer 11 days, and an invariant test would have caught it. The post Copilot Caught 8 of 8 Diff Bugs With Rules, 0 of 2 Outside the Diff…
For anyone running an inference server: a zero-variance control, five traffic traces, and an engine-vs-scheduler split show where the gains come from. The post Continuous Batching Beats Static and Dynamic by Up to 2.91× in vLLM appeared…
Studies of short stories, metaphors, and 486 participants point one way, an 800-person experiment points the other, and the author weighs both. The post Ask AI to Refine Your Ideas, Not Invent Them appeared first on Towards Data Science .
4 classic leaks, planted in real data, tested cold on 4 models.
A fisheries estimator from the 1890s, tested on 294 CUAD contracts, landed within 2 points of true recall for most agents with a 10% hand audit. The post How to Estimate the Recall of AI Agents with Capture-Recapture appeared first on…
Learn how to use Python scripts to automate repetitive tasks, organize files, process images, and simplify daily workflows.
On 88 house sales, OLS predicted $164K for a $219K home, while a 90% interval from 5th and 95th percentile models spanned $136K to $227K. The post Quantile Regression Caught the House Price OLS Missed by $55K appeared first on Towards Data…
Microservices changed how we build software by breaking large applications into smaller, independently deployable services. That made systems easier to develop and scale, but it introduced a different class of problems: The behavior of the…
Build an EV purchase adviser and a support agent in Python, then see how fan-out, pipelines, hierarchies, and swarms reuse the same building blocks. The post How to Build Multi-Agent Systems with the OpenAI Agents SDK appeared first on…
Even with step-by-step reasoning, the models landed near 33% accuracy, and one quoted the host's rule correctly before ignoring it. The post Change One Monty Hall Rule and Five LLMs Still Say Switch appeared first on Towards Data Science .
All four time-series foundation models scale, patch, and predict quantiles; the real split is how series share information. The post Timestral, t0, Tabby, and Morpheus-2.0 Compared on GIFT-Eval appeared first on Towards Data Science .
Bellwether joins support tickets, store reviews, sales calls, and NPS to CRM accounts, then stress-tests the ranking on 6,693 synthetic signals. The post How to Prioritize Product Feedback by Revenue at Risk appeared first on Towards Data…
Get unlimited free access to 365 Data Science’s complete AI and data learning platform, including 134 courses, hands-on projects, and certifications, through November 20.
New frontier models arrived this week as investors showed signs of growing cautious about AI spending. On the latest episode of This Week in AI, we examined those developments, along with research into how models learn and the less visible…
Includes a worked example with 50 users and $110,000 in first-year cost, plus baselines, targets, and ownership rules for every AI initiative. The post How to Measure AI Adoption, Performance, and ROI With 12 KPIs appeared first on Towards…