标签：#ics

Company noted for building quantum annealers now also making gate-based hardware.

Comparisons between SAFi and techniques such as reinforcement learning from human feedback, or RLHF, are useful only up to a point. Constitutional AI is a closer conceptual comparison because it introduces explicit principles into the process of generating and evaluating responses. Even so, these approaches address a different layer of the problem. RLHF and Constitutional AI primarily shape how a model behaves. SAFi governs how an AI agent operates. That distinction matters because an AI agent is not only a language model producing text. It may interpret requests, reason about possible responses, decide whether to act, call tools, access information, modify data, and produce an answer that must be accountable to the organization deploying it. The conventional architecture: the model at the center Much of today’s AI governance consists of filters, classifiers, guardrails, monitors, and policy checks placed around the model. The general pattern looks like this: A request reaches the model. The model generates a response or proposes an action. External controls inspect the input, output, or tool request. The system allows, blocks, modifies, or records the result. This architecture can be valuable. External controls can detect prohibited content, restrict certain actions, and provide monitoring or enforcement. They are often necessary parts of a responsible deployment. But the architecture still places the model at the center of the process. Governance is positioned around the model as an additional control mechanism. In many systems, the evidence needed for explanation and audit is also collected after the model has produced its output or proposed its action. That creates a basic separation between execution and governance: The model produces the draft. The governance system evaluates the draft. The monitoring system records what happened. The controls may be effective, but governance remains an external activity surrounding the primary intelligence. SAFi’s architectur

Joe Kent says that he and other anti–Donald Trump MAGA figures have formed a new movement. They want Tucker Carlson to run for the presidency in 2028.

TL;DR If you're a five-person startup shipping a Node.js API and you want a metrics dashboard by Friday, send your telemetry to a managed backend and keep only the instrumentation layer inside your own repo. The alternative — standing up a time-series database, an object store for long-term blocks, and a dashboard service — puts three more components on an on-call rotation that hasn't earned its first SLO yet. Settle the wire format now and treat the backend as a config line you can change later. I own the platform team's roadmap, which in practice means I'm the person who defends the monitoring bill in a budget review and also the person who gets paged when a disk fills at 03:00. Those two jobs pull in opposite directions, and most of the advice online is written by people who only hold one of them. Usually the pager wins the argument. Should a startup run its own metrics stack, or pay for a managed dashboard? Start with capacity, because that's the step everyone skips before signing anything. A moderately instrumented Node.js API — say 40 HTTP routes, two queue workers, default runtime and event-loop metrics, one latency histogram with ten buckets — sits somewhere around 3,000 to 8,000 active series per process. Multiply by replicas. Multiply again by every environment you keep alive, including the staging cluster nobody admits to. You are at 50k active series before a single engineer has written a custom counter, and a self-hosted scraper will chew through that on a 2 GB VM without noticing. It will still be fine at 500k. Past a few million active series you're into sharding, remote storage, and a retention argument with whoever pays for object storage — that's the point where the self-hosted route stops being free and turns into a project with a headcount attached. None of that work is hard. It's just never zero. Dimension Self-hosted stack Managed metrics backend Time to first dashboard 1–3 days under an hour Who owns retention you, plus the storage bill vendor

People type 2+2 into Google. They type 9+10 . They type 7*8 when they can't remember whether it's 54 or 56. Each of those is a real, high-volume search query — and most of the results are identical calculator widgets. So when I built Wrongulator , a calculator that returns a confidently wrong answer on purpose, I had a question worth asking: what if every expression were its own page, ranking for the exact arithmetic people already search? That is programmatic SEO — generating a page per parameter instead of writing pages by hand. And doing it across 17 languages means programmatic SEO with hreflang, where each generated page also declares its 16 translated siblings. The trap is that most programmatic surfaces are thin, duplicative, and get buried by Google. This one isn't, for a specific reason: every page has a real, unique answer baked into the HTML before any JavaScript runs. This post is about how — and the honest costs nobody mentions. Why a Permalink Per Expression Is Even Possible A page per expression only works if /2+2 reproduces the same result for everyone, forever, with no database behind it. That property isn't free — it's the result of one design decision I cover in detail in why a viral toy must be wrong the same way every time : the wrong answer is a pure function of the expression, seeded by a stable hash, with no per-user state. The relevant consequence here is what that property unlocks for SEO. Because f("2+2") always returns the same wrong answer, the server can compute that answer on demand for any expression in the URL, with zero storage. There's no pages table, no CMS, no pre-generation job. A request for /64+5 runs the engine, gets 67 ("the only correct number"), and renders a complete page around it. The programmatic surface is, in effect, infinite — but it costs nothing to hold, because nothing is stored. The pure function is what makes thousands of unique pages possible without a database. That's the foundation. Everything below is about

Part of "AI, engineering and what survives production", a series on the parts of building with AI that hold up once real traffic hits them. There is a claim going round that you have probably absorbed by now: AI-assisted development is making codebases worse. Refactoring is down, duplication is up, we are all writing more and revising less. The numbers behind it are real, the samples are enormous, and I found I had started repeating the conclusion in conversation without ever having checked it. Then it occurred to me that those figures are averages taken across hundreds of millions of changes from thousands of organisations, not one of which is mine. So what is the rate in your repository? Nobody has told you, and on current evidence nobody is going to. I set out to find mine, assumed it would take an afternoon, and spent three days discovering that the answer is far harder to get at than the confident version suggests. So this is not a piece about what AI does to code. It is about how to ask that question of your own repository without arriving at a wrong answer, which turned out to be the genuinely difficult part. The tool I built to do it is git-habits : free, local, and it reads no source code whatsoever. What git can actually tell you Git history is a surprisingly rich behavioural record. Not of quality, about which it knows nothing at all, but of habits: how often you commit, how large those commits are, whether you go back and change what you wrote last month, and whether anybody still touches the old code. That is a narrower thing than quality and it is the thing the industry claims has changed, so it is the thing worth measuring. Four signals are computable from commit metadata alone, without opening a single source file: Moved lines. The share of changed lines sitting in files git detected as renamed or copied. It is the closest thing history offers to "somebody went back and reorganised this." Legacy touch. The share of changes landing on files nobody has

In less than a week, Spider-Man: Brand New Day raked in $1 billion worldwide and had the biggest box office opening weekend in Hollywood history. The feature has been a reminder of why Sony is probably never going to give up the Spider-Man film rights, and highlighted how Marvel Studios was smart to strike a […]

The Trump administration shared the details of its plan with OpenAI, Anthropic, and other AI labs on Tuesday. For now, the public remains in the dark.

Quotient CEO Lizzie Matusov explains why soaring AI spend often fails to improve software delivery. She presents a research-backed AI maturity framework designed to help engineering leaders move beyond vanity metrics like token usage, align organizational AI adoption, and address critical bottlenecks across the software development life cycle to deliver measurable business outcomes. By Lizzie Matusov

Texas announced new a audit on data centers that could slow approval for new facilities seeking to connect to the state energy grid. Governor Greg Abbott (R) on Monday directed the Public Utility Commission of Texas (PUCT) and the Electric Reliability Council of Texas (ERCOT) to verify and audit new data center proposals, writing that […]

A portfolio can look good on the usual scorecard and still answer the wrong question. One line says return was high. Another says risk-adjusted performance was acceptable. A third says drawdown stayed inside a tolerable range. Then the market turns, the benchmark starts recovering, and the thing I actually care about is different: how efficiently did the portfolio catch up? That is where a new metric can fool its own author. If I build a recovery measure and it moves almost exactly like an existing ratio, I have created a longer name for the same signal. The right acceptance test is geometric: a useful metric should cast a different shadow. This is the rule I used while validating Hyperlogarithmic Benchmark Catch-Up Ratio (HBCR): orthogonality to existing measures is a first-class test, not a chart for the appendix. 1. A new metric has to earn its axis HBCR was built to measure benchmark-relative recovery dynamics. The research page states the motivation plainly: traditional benchmark-relative metrics often fail to capture the true dynamics of investment performance, especially during market recoveries [ A New Metric for Private Equity Risk Adjusted Returns , Calibration of Risk and Correlation in Private Equity ]. That framing matters because the obvious validation path is tempting and weak. You compare the new number with familiar performance measures, find a comforting relationship, and declare victory. But a high correlation with a well-known score can be a warning. If HBCR strongly tracked Sharpe Ratio, it would probably be an expensive synonym for risk-adjusted return. The acceptance test I wanted was sharper. HBCR should have some relationship with performance, because recovery has economic content. It should also avoid collapsing into the same direction as Sharpe Ratio, Beta, Volatility, Alpha, Total Return, or Max Drawdown. Written as a predicate, the test has two sides. Let $\mathcal{T}$ be the set of metrics already on the scorecard, $\rho_{n,m}$ the corr

Use metrics APIs for cron-job, API-failure, and business-event charts, then add a separate heartbeat monitor for jobs that never start. That is the smallest stack I would put on call for a small SaaS. A metrics dashboard can show success and failure counts, duration, backlog size, and error-rate trends; it cannot prove that a scheduler actually invoked a job. Healthchecks-style monitoring closes that specific gap. It still isn't full monitoring coverage, and I wouldn't describe it that way in an SLO review. The distinction matters because a failed run and a missing run leave different evidence. An API error usually increments something. A business event can be counted. A cron job that never fires may produce nothing at all — no duration, no failure, no final log line. No signal. How should a backend metrics dashboard combine cron jobs, API failures, and healthchecks? Start with the questions an operator must answer, not with a vendor menu. For cron jobs, I want a success count, a failure count, duration, and any queue backlog that can delay completion. For API failures, I want error counts and an error-rate trend beside request volume, because a raw count without a denominator can make ordinary traffic growth look like a regression. For business events, I want domain verbs: invoices issued, imports completed, or messages accepted. Those widgets belong on one dashboard because they describe the same service from different angles. Heartbeat monitoring is a separate control. A job reports a start or completion ping to Healthchecks, Cronitor, or an equivalent tool; if the expected ping doesn't arrive within its schedule and grace period, that system owns the missing-run signal. Keep that alert outside the metrics query path. Otherwise the component that failed to emit data is also the component being asked to notice its own silence. Silence counts. I've learned to write the failure matrix before drawing the dashboard. In one incident, a call returned 200, but the side e

Data pipeline automation has moved from a technical aspiration to a business imperative. As organisations deploy more AI agents that need fresh, reliable data, the volume and complexity of data pipelines has grown beyond what manual management can sustain. AI-powered pipeline automation — using AI to build, monitor, and repair data pipelines — is the emerging solution. Key Insight: AI-automated data pipelines reduce pipeline development time by 60%, decrease pipeline failures by 45%, and enable data teams to manage 3x more pipelines per engineer compared to manual approaches. The Pipeline Scaling Challenge The average enterprise now maintains 1,500+ data pipelines, according to research by Barracuda Networks, and this number is growing 25% annually as organisations add new data sources, new AI use cases, and new reporting requirements. Each pipeline has an average of 4.2 transformation steps, 2.1 data quality checks, and connects an average of 2.8 systems. The total pipeline infrastructure is complex, fragile, and increasingly beyond the capacity of manual management. The scaling challenge manifests in three ways. First, pipeline development backlog — the average data team has a 3-6 month backlog of pipeline requests from business users. Second, pipeline failures — the average enterprise experiences 15-20 pipeline failures per week, each requiring manual investigation and repair that consumes 30-40% of the data engineering team's capacity. Third, pipeline maintenance — as source systems change (schema updates, API modifications, deprecated fields), pipelines break silently and produce incorrect results until someone notices. This 'silent failure' problem is particularly dangerous because it erodes trust in data without anyone being aware that anything is wrong. The root cause of these challenges is that data pipelines have been built as static, manually-maintained infrastructure. A pipeline is coded, tested, and deployed. When the source or destination changes, a hu

The European Union has ushered in some additional rules that aim to make it easier for people to identify chatbots and AI deepfakes online. The new transparency obligations under the bloc's landmark AI Act came into effect on August 2nd, requiring companies to disclose when people are interacting with AI models, and if content has […]

Why do simple features suddenly require cross-team negotiations? In this article, explore how boundary drift quietly destroys change locality and increases cognitive load across teams. Learn practical sociotechnical strategies - redistributing mechanics, exposing essential policy, and rehearsing exception paths - to restore domain boundaries and enable a truly evolutionary software architecture. By Michael Fischer, Nicholas Lawrence, Monica Karekar

The ROI Black Hole in Social Marketing Consider a mid-market B2B software company whose social team manages campaigns across X, LinkedIn, Instagram, and TikTok from a single shared workspace. Each week the managers review platform-native dashboards that display rising follower counts, solid engagement rates on short-form video, and respectable click-throughs from carousel posts. They export weekly performance reports, paste the numbers into shared spreadsheets, and celebrate the month-over-month lift in impressions. Yet when the sales operations team asks which campaigns contributed to qualified pipeline, the social group cannot produce a single account-level match. Campaign links carry UTM strings, but many prospects arrive through mobile apps or shared links that strip those parameters, leaving the CRM with only anonymous referral domains and no usable journey data. The team attempts manual reconciliation by cross-referencing campaign dates with opportunity creation timestamps, but the exercise quickly collapses under volume. One campaign on LinkedIn might drive 400 clicks while another on TikTok drives 1,200, yet both appear in the CRM as undifferentiated social traffic. Without a consistent identifier that survives across platforms and into the marketing automation system, the social team cannot isolate which creative or audience segment produced the meetings that closed. Budget conversations therefore remain anchored to vanity metrics rather than incremental revenue, and executives grow increasingly skeptical of further platform spend. Medallion Architecture and the Absent Silver Layer Modern data platforms often organize information according to a medallion architecture that progresses through successive stages of refinement. The initial bronze layer captures raw event logs exactly as they arrive from each social API, preserving original timestamps, platform-specific identifiers, and unprocessed metadata. A subsequent silver layer then standardizes those recor

From how galaxies form to the beginnings of the universe, the images from the space telescope are giving cosmologists new data to explore—and theories to build out.

This guide is a collection or a summary on how to install and use a footprint of Apache Hadoop. I tried to follow an old version 2.7.1 guide that I created few years ago and adjusted this to use the latest version. Apache Hadoop 3.5.0 is used below; check the Apache releases page before future installations. These instructions target Linux (Ubuntu/Debian) for development or testing. Production clusters need Kerberos, network controls, encryption, monitoring, backups, and an upgrade plan. Do not expose HDFS or YARN ports to the internet. Native single-node installation Prerequisites sudo apt-get update sudo apt-get install -y openjdk-17-jdk openssh-client openssh-server pdsh curl tar java -version Hadoop requires Java and SSH; pdsh is recommended by the current Apache single-node documentation. Find JAVA_HOME if needed: readlink -f "$(command -v java)" | sed 's:/bin/java::' Download and install Pin the version for repeatable installs and verify Apache's SHA-512 checksum: export HADOOP_VERSION=3.5.0 cd /tmp curl -fLO "https://archive.apache.org/dist/hadoop/common/hadoop-${HADOOP_VERSION}/hadoop-${HADOOP_VERSION}.tar.gz" curl -fLO "https://archive.apache.org/dist/hadoop/common/hadoop-${HADOOP_VERSION}/hadoop-${HADOOP_VERSION}.tar.gz.sha512" sha512sum -c "hadoop-${HADOOP_VERSION}.tar.gz.sha512" sudo tar -xzf "hadoop-${HADOOP_VERSION}.tar.gz" -C /opt sudo ln -sfn "/opt/hadoop-${HADOOP_VERSION}" /opt/hadoop sudo chown -R "$USER":"$USER" "/opt/hadoop-${HADOOP_VERSION}" Add this to ~/.bashrc, adjusting JAVA_HOME if necessary: export JAVA_HOME=/usr/lib/jvm/java-17-openjdk-amd64 export HADOOP_HOME=/opt/hadoop export HADOOP_CONF_DIR="$HADOOP_HOME/etc/hadoop" export HADOOP_HDFS_HOME="$HADOOP_HOME" export HADOOP_YARN_HOME="$HADOOP_HOME" export HADOOP_MAPRED_HOME="$HADOOP_HOME" export PATH="$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin" Then load and verify it: source ~/.bashrc sed -i "s|^# export JAVA_HOME=.*|export JAVA_HOME=${JAVA_HOME}|" "$HADOOP_HOME/etc/hadoop/hadoop-env.sh" had

Every A/B testing tutorial ends the same way: run the test, wait for significance, ship the winner. Then you run a real test and variant B converts 12% better on newsletter signups, brings in 4% less revenue per visitor, and bounce is flat. Nothing is significant except the signups. Ship it? I spent an embarrassing amount of time on this question while building an A/B engine, and most of what I read online didn't help, because most of it assumes one metric. This post is what I ended up with. It's not novel — the statistics are decades old — but I couldn't find it written down in one place with working code, so here it is. Why the p-value doesn't answer the question you're asking Two problems, and the second one is the bad one. Multiple comparisons. Three metrics at α = 0.05 means roughly a 14% chance of at least one false positive if nothing is actually different. Bonferroni fixes this, but now you need α = 0.017 per metric and your test needs to run three times as long. On a site doing 300 conversions a month that's not a fix, it's a refusal. The p-value is answering a different question. It tells you the probability of your data assuming no difference exists. What you actually want to know is: if I ship B, how much do I expect to lose if I'm wrong? Those are not the same question and no amount of Bonferroni turns one into the other. There's also the peeking problem — everyone checks the dashboard daily and stops when it goes green, which quietly inflates the false positive rate well past whatever α you wrote down. I'll come back to that, because Bayesian methods do not magically solve it, whatever you may have read. Posterior first, decision second For a conversion rate, the Beta-Binomial conjugate pair gives you the posterior in one line. With a uniform prior, after c conversions out of n visitors: p | data ~ Beta ( 1 + c , 1 + n - c ) That's it. No closed-form comparison between two Betas that's worth implementing, so sample. PHP has no Beta sampler in core, and

The FBI, the EPA, and the Cybersecurity and Infrastructure Security Agency (CISA) have stopped short of officially blaming Iran for a spate of cyberattacks on Minnesota's water systems, but consensus is that Iran is likely behind them. That, of course, hasn't stopped Donald Trump from sharing his own theory - that Governor Tim Walz is […]