标签：#ci

The early website Hot or Not taught its users to rank people at scale. Twenty-six years later, dating apps are trying to escape the culture the site helped create.

Flock’s surveillance cameras have already sparked outrage. WIRED reconstructed its next-generation AI system, already in use by some police, to confirm it goes much further than tracking license plates.

Why plain k-fold silently overfits your trading model — and the 4-line fix that stops it. The Problem With k-Fold in Time Series Financial data is sequential. k-fold shuffles rows, so a training row from 2 PM Tuesday sits next to a test row from 10 AM Monday. Worse: triple-barrier labels overlap . A label at bar t looks 6 bars into the future; a training row at t+2 "knows" part of that future. The model leaks. V1's history is full of "HIGH overfit" verdicts — train AUC high, test AUC flat. Plain TimeSeriesSplit is only marginally better; it still lets adjacent windows bleed into each other. Purged + Embargoed CV For each test window [t0, t1] : Purge any train row whose label window overlaps the test window. Embargo max_training_horizon bars after the test window — drop those too. Overlapping labels are not i.i.d. Purging + embargoing makes the split honest. def purged_embargo_split ( n , n_splits = 5 , embargo_frac = 0.02 ): idx = np . arange ( n ) fold = np . array_split ( idx , n_splits ) splits = [] for i in range ( n_splits ): test = fold [ i ] emb = int ( len ( test ) * embargo_frac ) lo , hi = max ( 0 , test [ 0 ] - emb ), min ( n , test [ - 1 ] + emb + 1 ) train_mask = np . ones ( n , bool ); train_mask [ lo : hi ] = False splits . append (( idx [ train_mask ], test )) return splits Tune Only When You Have Enough Optuna once "won" a validation set with only 4 decisive rows — statistically meaningless. Rule: never tune when the decisive (non-abstained) validation rows are below ~30–50. Widen the date range or symbol basket first; don't trust the trial. Three-Way Split, Always train (fit) → validation (early stop + HP select) → disjoint calibration set (sigmoid/ isotonic) → test (untouched, final score only). V1 sometimes conflated validation and calibration. Keep them separate. The Promotion Gate Log every trial's train/val/test gap, not just the winner's test score. Promote only if replay AND shadow (≥1 live session) both beat baseline on buyer metrics : 1.5x

If you have ever copy-pasted a table out of a PDF, you already know what happens. Rows collapse into one long line of text. Columns interleave. Numbers land in the wrong cell, or no cell at all. The table on the page looks perfectly structured, but a PDF has no real concept of "table." It only knows where individual characters sit on a page. Every extraction tool, ours included, has to reconstruct the table from scratch, using nothing but the position of each word. That gap between "looks like a table" and "is structured data" is where almost every free PDF tool falls apart. Here is how we handle it, what actually works, and where it still doesn't. Two different jobs, two different tools PDFHaul splits this into two separate tools because they solve different problems. PDF to Excel rebuilds the whole document as a single spreadsheet, in the order it appears on the page: form labels, key-value pairs, section titles, and tables all together. It is for documents where you want the full content, not just the numbers, things like invoices, time sheets, and reports. Extract Tables does the opposite. It ignores everything that isn't a table and hands back one clean sheet per table, nothing else. It is for people who want structured data out, ready to sum, sort, and filter, not a copy of the document. Both tools share the same underlying geometry engine. The difference is what each one keeps and what it throws away. How Extract Tables actually decides what's a table The core problem with table extraction is that "looks tabular" and "is tabular" are not the same thing. A vector chart's axis box, a form's outlined signature field, and a two-column list of allergen names all produce something that a naive extractor will happily read as a grid. None of them are tables. Our pipeline handles this in four phases, all before anything is written to a spreadsheet: Phase 1: classify the page. Every page is scored as bordered (has ruled lines or filled-rectangle grid lines), stream (no

How the joule thief circuit “steals” energy from seemingly depleted power cells.

If rolled out, the feature would use the social media service’s TikTok Pay offering, which is already available in Southeast Asia for TikTok Shop purchases.

"A team of SpaceX engineers is on their way to conduct additional analysis on the vehicle."

Study finds X algorithm prioritizes ragebait that disproportionately impacts Democrats

Clear the Lineup Submission The Bug AI applications can fail silently — producing wrong outputs, degraded performance, or unexpected behaviors without explicit errors. In my case, the issue was SQL drift: queries executed successfully but returned incomplete or unstable results due to unsafe wildcard usage (SELECT *). This silent failure propagated downstream, degrading model accuracy without obvious alerts. The Fix I introduced an agentic validation and inspection layer into the pipeline using LangGraph, StatesGraph, MCP, and A2A. Inspection Layer: Deterministic checks (SQL linters, schema validators). Validation Layer: Agentic reasoning about query safety. MCP Integration: Standardized access to profilers and monitoring APIs. A2A Collaboration: Agents exchanged context to enforce compliance. This combination allowed the system to detect unsafe queries and route them for human review before deployment. PR Link Here’s the merged PR where the fix was implemented: Continental-Thaligai Repository – Merged PRs https://github.com/NikhilRaman12/Continental-Thaligai/pulse#opened-pull-requests Code Snippet python from langgraph import Graph from statesgraph import State from mcp import MCPClient class SQLInspection(State): def run(self, query): if "SELECT" in query and "*" in query: return {"risk": 0.7, "message": "Wildcard SELECT may cause drift"} return {"risk": 0.1, "message": "Query safe"} graph = Graph() graph.add_state("sql_inspection", SQLInspection()) graph.connect("sql_inspection", "human_review", condition=lambda r: r["risk"] > 0.5) result = graph.run("SELECT * FROM transactions") print(result) Diff Example: diff SELECT * FROM transactions SELECT transaction_id, amount, date FROM transactions This change eliminated silent drift in query results and improved reliability in downstream AI pipelines. Outcome Silent SQL drift eliminated. Improved accuracy in downstream AI models. Added regression tests to prevent recurrence. Strengthened CI/CD pipeline with agentic saf

The ChatGPT maker says its upcoming Astra model may have reached “critical” cyber capabilities, prompting it to halt a significant number of training runs while it tightens internal safeguards.

This is the latest large-scale DDoS attack to hit the social networking site this year.

The American automaker returns to the top category of endurance racing next year.

New study finds that the tusks have not one but two spirals, twisting in opposite directions.

Divine now has more than 2 million videos from Vine's original archive.

The universe may never tell you if your choices mattered. Owlcat’s Osiris Reborn might not either.

Flock em Wisconsin: por que 200 cidades derrubaram as câmeras colaborativas e o que fazer agora Introdução Em menos de um ano, a promessa de “vigilância democrática” da startup Flock virou manchete de retirada massiva: mais de 200 municípios de Wisconsin removeram os dispositivos instalados pelos próprios moradores. Falhas de privacidade, vazamento de imagens e retorno financeiro bem abaixo do esperado foram o gatilho de uma revolta que ainda ecoa nos fóruns de segurança pública. Este artigo prático mostra o que aconteceu , como a tecnologia funciona , quais foram os erros críticos e, principalmente, o que municípios, desenvolvedores e cidadãos podem fazer para evitar outro desastre . 1. Como a Flock operava (e onde estava o ponto fraco) Camada O que a Flock oferecia Problema crítico Hardware Câmeras IP de 1080p, custo médio US$ 45, instaladas em postes ou residências. Firmware aberto, sem assinatura digital – facilitava modificação mal‑intencionada. Conectividade Wi‑Fi ou rede celular 4G via SIM pré‑pago. Dados trafegados em HTTP sem TLS em 30 % das unidades. Armazenamento Cloud da própria Flock (AWS S3) com retenção de 90 dias. Criptografia “em repouso” apenas com chaves gerenciadas pela empresa; acesso interno ilimitado. Analytics IA de detecção de movimento e reconhecimento de objetos (pessoas, veículos). Modelo treinado com imagens públicas, mas sem anonimização – violava a Lei de Privacidade de Dados de 2022 (Wisconsin). Dashboard Portal web para gestores municipais visualizarem fluxos em tempo real. Falta de auditoria de logs; permissões de leitura concedidas a todos os usuários do portal. Exemplo de configuração vulnerável (arquivo flock_camera.conf ) # Configuração padrão enviada a todos os dispositivos [network] wifi_ssid = "FlockPublic" wifi_password = "default123" ; senha padrão nunca alterada [security] tls_enabled = false ; HTTP puro auth_token = "abcdef123456" ; token fixo, reutilizado em milhares de câmeras [storage] bucket = "flock-data-wi" encrypti

Brad Grantham discusses how software engineers and architects can transition from individual contributors to influential technical leaders. Brad shares actionable insights on expanding skills into business and legal domains, adapting communication styles for non-technical stakeholders, moving past ego to empower teams, and navigating complex organizational dynamics to maximize engineering impact. By Brad Grantham

Research shows heat amplifies the dangers of pesticides.

AI companies like Anthropic and OpenAI regularly publish reports on how people are using products like Claude and ChatGPT, but they only release the data they want us to see, AI researchers say. “There is no independent source to corroborate it,” says Anka Reuel, a Computer Science PhD candidate at the Stanford Trustworthy AI Research…

Successfully onboarding an automated test suite to Harness CI requires configuring infrastructure placeholders, secrets, pipelines, and branch protection rules. Here is a 10-step checklist to help you onboard your end-to-end (E2E) automation pipelines seamlessly. Step 1: Replace Infrastructure Placeholders Ensure your pipeline YAML definitions (e.g., .harness/e2e-poc.yaml and .harness/e2e-regression-parallel.yaml) contain your specific environment values: ORG_ID: Harness Organization Identifier PROJECT_ID: Harness Project Identifier GIT_CONNECTOR: Harness Git Connector for GitHub Enterprise access APP_REPO_NAME: Target repository in owner/repo format K8S_CONNECTOR: Kubernetes connector for build infrastructure K8S_NAMESPACE: Kubernetes namespace where build pods run Step 2: Configure Environment Secrets In Harness, set up the following runtime secrets: CONNECT_URL CONNECT_USERNAME CONNECT_PASSWORD Step 3: Setup PR Validation Pipeline Import your short-run pipeline YAML into Harness. Save it as your PR Validation Pipeline. Run a manual validation test using runtime overrides: TargetEnv = qa cucumberTags = @smoke Step 4: Verify Artifact Generation Confirm that the initial execution correctly generates and uploads all required outputs: JUnit Report: reports/junit-report.xml Test Reports: reports/** Failure Artifacts: test-results/** (screenshots, traces) Step 5: Setup Nightly Parallel Pipeline Import your parallel pipeline YAML into Harness. Save it as your Nightly Regression Pipeline. Run a manual validation test with target concurrency parameters: TargetEnv = qa cucumberTags = @regression cucumberParallel = 4 Step 6: Configure Automated Triggers & Branch Protection PR Trigger: Configured on pull requests with cucumberTags= @smoke . Nightly Schedule Trigger: Configured on a nightly cron schedule with cucumberTags=@regression and cucumberParallel=4. GitHub Branch Protection: Enable branch protection on target branches requiring the Harness PR pipeline status check to p