Close Menu
    Facebook LinkedIn YouTube WhatsApp X (Twitter) Pinterest
    Trending
    • These Were My Favorite Things Samsung Unpacked During Its 2026 Galaxy Event
    • AI minister role boosted but tech department axed in Burnham shake-up
    • Loop Engineering for RAG Question Parsing: The Small Loop That Runs Before Retrieval
    • The risk of weather data sabotage is rising
    • Hand-E Now Reaches 100 mm Without Giving Up an Ounce of Precision
    • Weight loss drug effectiveness and long term maintenance
    • Here’s what Albo’s ‘Office of AI’ means for Australian tech
    • YouTube and X Have Become ‘Gateways’ to Nudify Apps
    Facebook LinkedIn WhatsApp
    Times FeaturedTimes Featured
    Thursday, July 23
    • Home
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    • More
      • AI
      • Robotics
      • Industries
      • Global
    Times FeaturedTimes Featured
    Home»Artificial Intelligence»The Arithmetic of Productivity Boosts: Why Does a “40% Increase in Productivity” Never Actually Work?
    Artificial Intelligence

    The Arithmetic of Productivity Boosts: Why Does a “40% Increase in Productivity” Never Actually Work?

    Editor Times FeaturedBy Editor Times FeaturedApril 7, 2026No Comments6 Mins Read
    Facebook Twitter Pinterest Telegram LinkedIn Tumblr WhatsApp Email
    Share
    Facebook Twitter LinkedIn Pinterest Telegram Email WhatsApp Copy Link


    Introduction: False guarantees?

    as a advisor and supervisor within the information sphere, I’ve sat by means of my fair proportion of slide deck presentation. On each side. And any slide deck value its salt guarantees one thing, typically about effectivity or productiveness. You could have most likely heard one thing like this:

    • This device with make your information scientists 40% extra productive!
    • You’ll spend 30% much less time fixing bugs by doing this. You may mainly implement a 6-hour work day and nonetheless come out on prime!
    • With our answer, you may code up two initiatives within the time it took you beforehand to do just one. This halves the period of time to manufacturing!

    Generally the guarantees don’t work just because the proposed product is dangerous. However why does this by no means appear to really work even with good merchandise? You may find yourself switching to a product that you just genuinely love, however nonetheless not likely see the promised enchancment. Why? Are the numbers you bought offered lies?

    My background as a PhD in arithmetic has most likely scarred me for all times in additional methods than one. One of many deepest scars is my want to know exactly what numbers signify. And the numbers you discover within the statements above are all indicating one factor, whereas telling a totally totally different story when you cease and assume.

    Whereas mendacity actually occurs, what’s a way more frequent observe is being deceptive. This kind of advertising assumes that you just don’t assume critically when offered with numbers. Let’s assume critically collectively and see what we give you.


    Lies, lies, and advertising

    So what’s the concern with the productiveness statements?

    The primary concern is that they declare to optimize a sure side of the method, whereas (not directly) promising world productiveness good points.

    Let’s undergo a easy instance to know what this implies.

    Say you’re an enormous participant in AI and have lately launched a product that’s nice at serving to information scientists with mannequin parameter choice. Cool! Preliminary surveys present that it has given information scientists a 20% enhance in productiveness for mannequin parameter choice. You initially current this as:

    Our device has improved the productiveness of mannequin parameter choice for information scientists by 20%.

    Proud of this spectacular outcome, you ship your assertion of to advertising and so they come again solely with minor changes:

    Our device has improved mannequin parameter choice, making information scientists 20% extra productive.

    You shrug and surprise for a second what these folks in advertising are actually paid to do once they solely shuffled round just a few phrases. In actuality, they’ve now shifted your assertion from one thing that’s reasonably spectacular to one thing that’s insanely spectacular.

    Why? The adjustment from advertising makes it appear to be the product makes information scientists 20% extra productive normally. However your survey solely actually talked about productiveness in the course of the time the information scientists are deciding on mannequin parameters. What’s the true distinction?

    An information scientist does a great deal of issues, together with prototyping, stakeholder administration, coordination conferences, and many others. Whereas machine studying is usually on the entrance and heart of how they’d describe themselves, many information scientists solely spend round 40% of their time on typical information science duties. A giant chunk of those 40% is debugging information high quality points, pipeline administration, and information validation. Mannequin parameter choice might solely take up 10% of their 40% time doing information science duties. Multiplying reveals us that that is solely 4% of their complete time.

    If the information scientist adopted a device to make the mannequin parameter choice 20% extra productive, that might solely make a distinction of barely 1% of their complete time. You wouldn’t discover this throughout a piece week. In actual fact, with the added complexity of studying a brand new device at first, you may even see a lower in productiveness at first.

    One of the best half? Take a look at the assertion rigorously:

    Our device has improved mannequin parameter choice, making information scientists 20% extra productive.

    It actually appears to argue that information scientists could be 20% extra productive general, however this is only one interpretation. If pressed, advertising would make the connection between the beginning and finish of the sentence, and say that it’s implied that the productiveness enhance is just for mannequin parameter choice.

    So that you successfully get to say one factor, whereas falling again on one other if the deceptive assertion is found. The pay for advertising comes from shuffling across the proper phrases!


    A greater method? Deal with cognitive load relatively than productiveness

    What does the story I simply informed actually let you know? You probably have many various duties which can be advanced (like a knowledge scientist does), then aiming for productiveness good points is actually not pushing the needle that a lot.

    Don’t get me flawed. You probably have a straightforward alternative to grow to be 20% extra productive with one among your duties, go for it! However don’t anticipate it to end in greater than a p.c or two distinction in complete productiveness.

    What can we do as an alternative when we have now many various duties which can be advanced? We will use cognitive load as a metric, and attempt to scale back that as an alternative.

    Say {that a} competing firm developed their very own device for mannequin parameter choice. As a substitute of attempting to hurry up the method, their device had the only real objective of decreasing the cognitive load of the information scientist. So the method of mannequin choice would take the identical period of time, however the information scientist would really feel energized and prepared for an additional problem after deciding on mannequin parameters.

    Most individuals, myself included, can’t work 8 hours a day and be on the prime of our sport always. Some days I really feel like I’ve 6 efficient hours in me. Different days it’s extra like 2 efficient hours. If one course of doesn’t require that a lot cognitive load, then I can work longer successfully. This typically leads to the identical complete productiveness of some p.c, however with the additional advantage of improved morale.

    So subsequent time somebody presents a “40% enhance in productiveness“, ask them the next:

    • How a lot of the entire work time does this productiveness enhance have an effect on?
    • How a lot cognitive load does this take away or introduce?



    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Editor Times Featured
    • Website

    Related Posts

    Loop Engineering for RAG Question Parsing: The Small Loop That Runs Before Retrieval

    July 19, 2026

    How to Find the Optimal Coding Agent Interface

    July 9, 2026

    I Completed Five Years in Analytics Consulting: 5 Lessons That Changed How I Work

    June 29, 2026

    GPU-Resident Top-K for Agentic RAG: I Built a CUDA Kernel So My Retrieval Step Would Stop Bouncing Off the GPU

    June 19, 2026

    Can Machine Learning Predict the World Cup?

    June 9, 2026

    Automate Writing Your LLM Prompts

    June 5, 2026

    Comments are closed.

    Editors Picks

    These Were My Favorite Things Samsung Unpacked During Its 2026 Galaxy Event

    July 22, 2026

    AI minister role boosted but tech department axed in Burnham shake-up

    July 21, 2026

    Loop Engineering for RAG Question Parsing: The Small Loop That Runs Before Retrieval

    July 19, 2026

    The risk of weather data sabotage is rising

    July 18, 2026
    Categories
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    About Us
    About Us

    Welcome to Times Featured, an AI-driven entrepreneurship growth engine that is transforming the future of work, bridging the digital divide and encouraging younger community inclusion in the 4th Industrial Revolution, and nurturing new market leaders.

    Empowering the growth of profiles, leaders, entrepreneurs businesses, and startups on international landscape.

    Asia-Middle East-Europe-North America-Australia-Africa

    Facebook LinkedIn WhatsApp
    Featured Picks

    The Pandemic Appears to Have Accelerated Brain Aging, Even in People Who Never Got Covid

    July 30, 2025

    Optimizing Data Transfer in AI/ML Workloads

    January 3, 2026

    Today’s NYT Strands Hints, Answer and Help for May 26 #814

    May 26, 2026
    Categories
    • Founders
    • Startups
    • Technology
    • Profiles
    • Entrepreneurs
    • Leaders
    • Students
    • VC Funds
    Copyright © 2024 Timesfeatured.com IP Limited. All Rights.
    • Privacy Policy
    • Disclaimer
    • Terms and Conditions
    • About us
    • Contact us

    Type above and press Enter to search. Press Esc to cancel.