Artificial Intelligence’s Data Needs: Can They Be Met Legally and Logistically?

Three of the problems I identified with AI in previous posts concern the data used to train its large language models.  One is the sheer volume of information it needs to create more advanced capabilities.  Second is data’s legal status, which has caused several large lawsuits, and doubtless many more small ones, charging copyright infringement.  The third is distortion from chatbots taking in output from themselves or others.  What has been in the press lately about these issues, and what does it mean not only about this aspect of AI but about AI in general?

Apparently, “Apple, Nvidia, Anthropic Used Thousands of Swiped YouTube Videos to Train AI” (Annie Gilbertson, WIRED, July 16^th).  The problem has been that “tech companies are turning to controversial tactics to feed their data-hungry artificial intelligence models, vacuuming up books, websites, photos, and social media posts, often unbeknownst to their creators.”  Everyone anywhere near the field, let alone companies’ legal personnel, should know that electronic versions of books and published articles are as subject to copyright laws as hardcopy editions, long documented in statements such as “no part of this book may be reproduced in any form or by any means without the prior written permission of the Publisher, excepting brief quotes…” which I got from a random 1968 paperback – but it is understandable for lay people not to know if that also applies to the likes of videos and other less formally protected online material.  It also may be difficult, in these data-absorbing efforts, to avoid off-limits products, but the problem still must be solved.

That’s why, at least per Nico Grant and Cade Metz, in the New York Times on July 19th, we are seeing or should see “The Push to Develop Generative A.I. Without All the Lawsuits.” The partial copyrighted-information solution here is those owning the rights to data “building A.I. image generators with their own data,” and then selling AI-development access.  Two companies already starting that are “the major stock photo suppliers Getty Images and Shutterstock,” which will pay photographers when their work is thus used.  Fair play, or so it seems.

Otherwise, “The Data That Powers A.I. Is Disappearing Fast” (Kevin Roose, The New York Times, July 19^th).  Although, per research “by the Data Provenance Initiative, an M.I.T.-led research group,” “three commonly used A.I. training data sets” had restricted only 5% of their data (though “25 percent… from the highest-quality sources”), but the operation is in progress.  Conclusive definition of legal information use is not here yet, as “A.I. companies have claimed that their use of public web data is legally protected under fair use.”  Perhaps, per the author, “if you take advantage of the web, the web will start shutting its doors.”

Another way out was described in Forbes Daily on July 24^th: “The Internet Isn’t Big Enough To Train AI.  One Fix?  Fake Data.”  “OpenAI’s ChatGPT, the chatbot that helped mainstream AI, has already been trained on the entire public internet, roughly 300 billion words including all of Wikipedia and Reddit” (italics in original), meaning that “at some point, there will be nothing left.”  A company, Gretel, wants to provide AI firms with “fake data made from scratch,” which is not totally new, as “Anthropic, Meta, Microsoft and Google have all used synthetic data in some capacity to train their models.”  Two issues with it are that “it can exaggerate biases in an original dataset and fail to include outliers,” which “could make AI’s tendency to hallucinate even worse.”  If, that is, it does not “simply fail to produce anything new.”  We will find out, probably within the year, if artificial data is a worthwhile partial or complete substitute.

To the point of the final first-paragraph problem is “What happens when you feed AI-generated content back into an AI model?  Put simply:  absolute chaos” (Maggie Harrison Dupre, Futurism.com, July 26^th).  Per a recent study, “AI models trained on AI-generated material will experience rapid “model collapse” … as an AI model cannibalizes AI-generated data, its outputs become increasingly bizarre, garbled, and nonsensical.”  The problem is out there now, as “there are thousands of AI-powered spammy “news” sites cropping up in Google; Facebook is quickly filling with bizarre AI imagery… Very little of this content is marked as AI-generated, meaning that web scraping, should AI companies continue to attempt to gather their data from the digital wilds, is  becoming a progressively dubious means of collecting AI training data.”

Despite the hope in the second story above, none of this looks good for future AI releases.  These problems will not be easy to solve.  We already have the issue that AI is nowhere near ready to produce even page-length writing releasable without human scrutiny – the concerns here will, most likely, keep that capability at bay.  Until then, AI will fail to even approximate the utility expected by its customers and backers.  That means, even without regard to other obstacles such as insufficient power for fundamentally more advanced releases, that artificial intelligence is in deep trouble.  All should govern themselves accordingly.

Artificial Intelligence Investments and Revenue – A Month’s Worth

AI, you’ve done it to me.

Even when eliminating most of the news items about it, those focusing on the future instead of the present and on other factors irrelevant to what is happening with the technology itself, there are too many for even weekly posts.  So I have divided it into subtopics, of which this is the first to get publication. 

The oldest here is “Microsoft’s Steep AI Investments Raise Questions About Returns” (Patrick Seitz, Investor’s Business Daily, July 10^th).  As crisply put by a Morgan Stanley analyst, “with Microsoft’s capital expenditures poised to nearly double from $32 billion in fiscal 2023 to our forecasted $63 billion in fiscal 2025, the question of monetization against these investments rises to the forefront for many investors.”  The analyst still rated the stock highly, but, per the author, “after digging into the question of return on investment,” “was left with more questions than answers” because of a “lack of visibility.”

The first item in the July 19^th Goldman Sachs Briefings was “Will the $1 trillion in AI spending pay off?.”  The determination here was “AI may prove far less promising than many business leaders and investors expect,” per an MIT professor who “estimates that only a quarter of AI-exposed tasks will be cost-effective to automate within the next 10 years – which means that AI will impact less than 5% of all tasks and boost US productivity by only 0.5% and GDP growth by 0.9% cumulatively over the next decade.”  A Goldman Sachs research head went “a step further” by saying that “there’s not a single thing that this is being used for that’s cost-effective at this point.”

On the other hand, “Alphabet Reports 20% Jump in Profit as A.I. Efforts Begin to Pay Off” (Nico Grant, The New York Times, July 23^rd).  “The company has incorporated generative A.I. into all of its products and increased its spending on data centers and associated hardware to underpin the technology.  Now, Google executives said, those investments have started to bear fruit,” with second-quarter profit of $23.6 billion.  However, Victor Tangermann, in “Investors Are Suddenly Getting Very Concerned That AI Isn’t Making Any Serious Money” in Futurism.com on July 27^th, reported that “investment bankers are… starting to become wary of Big Tech’s ability to actually turn the tech into a profitable business," saying that Google’s earning report was “failing to impress investors with razor-thin profit margins and surging costs related to training AI models.”  Still pertinently, a “tech stock analyst” wrote back in March that “capital continues to pour into the AI sector with very little attention being paid to company fundamentals… a sure sign that when the music stops there will not be many chairs available.” 

Perhaps in reaction to these pieces and the thoughts behind them, “Tech Bosses Preach Patience as They Spend and Spend on A.I.” (Karen Weise, The New York Times, August 2^nd).  “The tech industry’s biggest companies have made it clear over the last week that they have no intention of throttling their stunning levels of spending on artificial intelligence, even though investors are getting worried that a big payoff is further down the line than once thought.”  On a week ago Thursday, Mark Zuckerberg of Meta said he would spend “at least $37 billion” on “new tech infrastructure,” and “would spend even more next year.”  “In the last quarter alone, Apple, Amazon, Meta, Microsoft, and Google’s parent company Alphabet spent a combined $59 billion on capital expenses, 63 percent more than a year earlier and 161 percent more than four years ago.”  Another Goldman Sachs executive asked “What $1 trillion problem with A.I. solve?,” and implied that “replacing low wage jobs with tremendously costly technology” was not sufficient justification.  Yet leaders of the other largest IT firms also emphasized the need to be at the front of AI progress, and that true business success there might take a long time.

Is it true that “The AI supply chain is in jeopardy” (Guy Scriven, The Economist, August 3^rd)?  “About a year ago,” he “started to ask analysts what it would take to stop the… (AI) bull run.  Most suggested it would end if big tech firms failed to deliver meaningful AI-related sales for a couple of quarters in a row.  It took more than a couple of quarters, but over the past two weeks, as the tech giants reported their earnings, that scenario has started to play out.”  Adding to general murkiness is that Microsoft, where “growth in this revenue stream has been slowing,” “is still the only big tech firm that puts a figure to its AI sales.”  He echoed that “scepticism has set in among investors,” and opined that “the boom looks somewhat precarious,” as “the larger risk is that demand peters out.”

Will it, or won’t it?

July Jobs Report: Downward Trend Continues – AJSN 600,000 Higher, With Latent Demand Over 17.8 Million

 I saw three published estimates of the number of net new nonfarm payroll positions in this morning’s Bureau of Labor Statistics Employment Situation Summary.  All three were 175,000. 

It didn’t make it – the number was 114,000.  That’s still decent, as American population increased only 163,000, but was the worst in months. Even poorer in relation to recent outcomes was the seasonally adjusted unemployment rate, up from 4.1% to 4.3%.  Other results inferior to June’s were unadjusted joblessness at 4.5% instead of 4.3%, the number of unemployed up 400,000 to 7.2 million, the employment-population ratio from 60.1% to 60.0%, and the count of those working part-time for economic reasons, or thus far unsuccessfully seeking full-time positions while keeping shorter-hours ones, up a disturbing 400,000 to 4.6 million.  Average hourly nonfarm private payroll earnings again roughly matched inflation, up 8 cents to $35.07, and the number of long-term unemployed, those jobless for 27 weeks or longer, remained at 1.5 million.  On the good side, the number claiming no interest in work lost 804,000 for a 1.41 million drop over two months, to reach 92,972,000, and the labor force participation rate gained 0.1% to get to 62.7%.

The American Job Shortage Number or AJSN, the measure showing how many additional positions could be quickly filled if all knew they would be as easy to get as a lottery ticket, gained 602,000 as follows:

The share of the AJSN from those officially jobless was 38.7%, up from 37.7%.  Compared with a year before, the metric is almost 1.2 million higher, almost matching the difference from more unemployment but also reflecting gains in those discouraged and those wanting work but not looking for it for a year or more. 

Some jobs reports are confusing and contradictory, but this one was not.  People are still joining the labor force in large numbers, shown by the 1.41 million above and by the count of those employed actually going up, 264,000 to 162,038,000.  When you look at the discouraged and did-not search outcomes, along with the unemployment results, you can see that many Americans are either losing work or not getting it.  The labor force is growing faster than the number of positions.  Inflation is now at 2.5%, reasonable in every way except in relation to the Federal Reserve’s target, but we clearly need more jobs – especially full-time jobs.  The Fed missed a chance to cut interest rates this week; at next month’s meeting, they shouldn’t, and a nominal quarter point, unless the data above substantially improves, will not be enough.  For now, the turtle stayed right where he was.

Electric Vehicles are Going Nowhere Special

In my December 15^th post, I wrote “fundamental sale price drops may require more to be sold than the market can support.  I predict that by November 2024 the share of electrics as American vehicles will be higher than it is now, but only about 10%, with a growing consensus that under current conditions they will not be completely or even largely taking over.”  Seven-plus months out, how are those forecasts looking?

On May 26^th in Fox Business, we saw how “Buttigieg defends Biden’s EV strategy after question on how only 8 federal charging stations have been built.”  The unfortunately-placed transportation secretary was asked why, when two years ago President Biden signed a bill authorizing 500,000 such facilities to be built by 2030 and after one-quarter of the time since has elapsed, the number completed stands at .0016% of that; he reiterated the half-million goal and said there was “utility work” to be done at each, under “a new category of federal investment.”  How many have been started?  I could find information on ports, overall stations, and groups of stations with at least one under construction, but not that.

Another thing Buttigieg mentioned in the interview above was the necessity of lower electric vehicle prices.  Soon afterward, on June 3^rd, the New York Times published “Electric Cars Are Suddenly Becoming Affordable” (Jack Ewing).”  The reason, though, was not higher sales providing manufacturing economies of scale, but that “customers have been snapping up used Teslas for a little over $20,000, after applying a $4,000 federal tax credit.”  That trend started with Hertz’s January-announced sale of 20,000 electric rental cars, which the market has not yet absorbed.  Ewing also mentioned, for new cars, the effect of “increased competition, lower raw-material costs and more efficient manufacturing,” but companies choosing to cut prices were pushed by Hertz, as “electric cars still cost more to manufacture than cars with internal combustion engines,” and sales would further drop without continuing subsidies.

The lack of growth has also pushed companies to build “More Gas Cars and Trucks, Fewer E.V.s as Automakers Change Plans” (Neal E. Boudette, The New York Times, July 18^th).  Ford Motor Company “said it would retool a plant in Canada to produce large pickup trucks rather than the electric sport-utility vehicles it had previously planned to make there,” and the day before that, “General Motors said it expected to make 200,000 to 250,000 battery-powered cars and trucks this year, about 50,000 fewer than it had previously forecast.”  Tesla, as well, “has changed its plans because it no longer expects sales to grow 50 percent a year,” as “its global sales fell 6.6 percent in the first six months of the year.” 

And speaking of the last company mentioned, “Tesla’s Profit Fell 45% in the Second Quarter on Weak E.V. Sales” (Neal E. Boudette, The New York Times, July 23^rd).  Its second-quarter year-over-year revenue increased from $24.9 billion to $25.5 billion, but its net profit fell from $2.7 billion to $1.5 billion.  That seems to confirm that its shortfall from lower selling prices is not from lower costs.  For the same interval, its number of EV’s sold dropped 4.8% and their production 14%. 

As of an Experian Automotive Market Trends fourth quarter 2023 report, only 3.3 million out of 288.8 million American vehicles were electric.  Per Edmunds, 6.9% of sales from January to May this year were the same.  Therefore, it is almost impossible to imagine the year-end share will be anywhere near 10%, let alone higher.  It looks also very doubtful that, given the news above, EV’s will be “completely or even largely taking over” – even if a Democrat wins the November presidential election.  Their percentages may substantially increase, but until then, we must project electric vehicles, with some geographical exceptions, to remain a small minority indefinitely.  If their manufacturers cut back and switch to others, and subsidies continue to prop them up, it will mean little if those in the press say they will predominate.  The lack of federal charging stations may, in the end, be just fine, and customers will get the vehicles they want.  That will work.

Four Areas of the Economy – Four Views on Them

What American economic segments have been analyzed over the past four months?

One thing dragging us down over the past several years is tariffs.  About that, David Wallace-Wells asked, in the New York Times on May 22^nd, “Can Biden Actually Win This Trade War?”  I mentioned this massive increase, which was actually “a unilateral quadrupling” on Chinese electric vehicle duties, on a post about EV’s.  It seemed destructive for a president who seemed to badly want Americans to get away from gasoline and diesel cars, and, per the author here, “it’s not just EVs.  Five years after blasting (then-President Donald) Trump for imposing tariffs on Chinese exports, Biden raised them – on aluminum, steel, lithium batteries, solar cells and semiconductors, among other products.”  The damage to other environmentally favorable efforts is also great, as China’s production shares of related items include “84 percent of the world’s solar modules… 89 percent of the world's solar cells… 97 percent of its solar wafers and ingots, 86 percent each of its polysilicon and battery cells, 87 percent of its battery cathodes, 96 percent of its battery anodes, 91 percent of its battery electrodes and 85 percent of its battery separators.”  Although various “Democratic policymakers” say we should “avoid growing entirely dependent on China for clean energy,” “industrial policy isn’t guaranteed to work,” and “Trump’s imposition of tariffs on Chinese solar-panel exports in 2018 may have meaningfully slowed American renewable rollout.”  The extra charges have real potential to continue doing that.

Something around for many decades and worsening every year gets concern of various strengths every so often, but there are reasons, per Paul Krugman on June 6^th in the New York Times, “Why You Shouldn’t Obsess About the National Debt.”  Although, per usdebtclock.org, it is now almost $35 trillion, or over $100,000 per citizen, Krugman calls it “a lot less scary than many imagine if you put it in historical and international context,” and making it “sustainable wouldn’t be at all hard in terms of the straight economics; it’s almost entirely a political problem.”  As a share of GDP, the author said “it’s roughly the same as it was at the end of World War II,” and is now “considerably lower” than Japan’s.  We could “stabilize debt as a percentage of G.D.P. for the next 30 years” by raising taxes or reducing spending only 2.1%.  So, the money we owe is real, but it may be dormant instead of a looming threat.

Do we have “A towering real estate crisis” (Andrew Ross Sorkin, The New York Times DealBook, June 12^th)?  “The $2.4 trillion office building sector” has been hit by “sky-high interest rates and a pandemic-induced shift to remote work,” as “tenants are reducing or vacating office space to landlords at a record clip.”  As a result, “developers are looking to repurpose empty offices,” although “conversions are expensive, and not all buildings can be easily retrofitted.”  This problem may be worse than it seems, as the pendulum I have documented which swings back and forth between businesses favoring remote and in-office work is approaching the latter maximum, so it will be almost one complete cycle before office demand increases substantially, if it does at all.  Accordingly, a surplus of commercial space will seem to this generation like a permanent problem, so it needs permanent solutions.

I’m used to issuing the AJSN and basing my views using information I trust, so was not happy to see “Reliability of U.S. Economic Data Is in Jeopardy, Study Finds” (Ben Casselman, The New York Times, July 9^th).  Per an American Statistical Association report, while “government statistics are reliable right now,” “that could soon change,” due to “shrinking budgets, falling survey response rates and the potential for political interference.”  Responses to the Current Population Survey, one of two providing the foundation for “the unemployment rate and other labor force statistics,” have dropped from “nearly 90 percent” ten years ago to “about 70 percent in recent months,” and those to “other government surveys” are down also.  As for the third problem, “there are few legal provisions ensuring that the statistical agencies can operate independently,” and an author of the report called for ““legislation to make this issue of professional autonomy statutory.””  We will see, and weakness here will depend on the result of the upcoming presidential election.  The same will affect the previous three items as well – along with a great deal more. 

A Directionless Month for AI – What Do Its Scattered Reports Say?

Artificial intelligence seems to be moving into another phase, or subphase.  There have never been so many concerns, not about the technology causing a major disaster, but about it being adequate and worthwhile at all.  At the same time, there are new perceptions and more progress in related employment, and, if I am perceiving it correctly, growing acceptance that where AI is now might be where it is for months or even years, so stakeholders are realizing they cannot just wait for the next big eruption. 

A look at one chronic AI problem is the subject of “How Game Theory Can Make AI More Reliable” (Steve Nadis, Wired.com, June 9^th).  The author bemoaned the tendency of large language models to give different answers, not all correct, to generative (open-ended) and discriminative (choice of options) questions.  Researchers have invented a game of sorts where “two modes” of LLMs are asked to agree on answers.  Excellence at games is a long-time AI strength, so engineers have reason to be optimistic such inconsistencies will go away.

Top AI workers are scarce, and many, soon after starting work with an organization, move on to another for more pay.  That is why, on one level, “Retention is all you need” (The Economist, June 15^th).  As, per a market analyst, “20,000 companies in the West are hiring AI experts,” “AI talent, previously hoarded at tech giants, is becoming more distributed.”  Indeed, AI hiring at Amazon, Apple, Google, Meta, and Microsoft has only about broken even since the beginning of 2023.  Advertisements for many more software positions now mention AI, as do those for math, scientific research and development, and information design.

Defining another clear problem, Sherin Shibu’s “How Can AI Help Small Businesses?  It’s A Matter of Trust, According to a New Report” (Entrepreneur, June 17^th), tells us that “just 7% of U.S. desk workers see AI answers as completely trustworthy,” the share of American businesses as of June was only 5% although 96% of “surveyed executives felt pressure to bring AI into their business,” and “privacy” and “data quality” were the other two top issues.  These tell AI companies where they should focus, perhaps above all other considerations.

Unsurprisingly, more universities and technical schools are preparing students for jobs using the technology.  A Brock Dumas June 17^th Fox Business piece, “Want an AI career?  These colleges offer degrees for the best chance” provides a current look.  “A new study by software development firm Vention shows which U.S. colleges are offering bachelor’s degrees with the best chance of landing a job fresh out of school – and the potential for the biggest paychecks.”  The top five were California Polytechnic State University, Wake Forest University in North Carolina, Trinity University in Texas, Clarkson University in New York state, and Knox College in Illinois.  There are more, and this list will change, but these may be the ones currently most worth investigating.

Also, for now, “We’re Still Waiting for the Next Big Leap in AI” (Will Knight, Wired.com, June 20^th).  Knight wrote that “the world is still waiting for another AI leap forward… akin to that delivered by GPT-4” 16 months ago.  He described possible contenders, but then concluded “it’s unclear how long the world must wait for that next big leap in AI.  OpenAI has said it has started training its next big model.  In the meantime, we will need to figure out new ways to measure how useful the technology really is.”

On June 27^th, Fox News released “US tops world ratings for AI preparedness:  China, Russia and Iran lag in key measures, report finds” (Peter Aitken).  The rankings were on countries’ “ability to immediately adopt artificial intelligence… into their economies.”  The American “value of preparedness” tied the Netherlands, and was followed by Finland, Estonia, New Zealand, Germany, Sweden, Australia, Japan, and Israel.  That may or may not prove to be valuable information.

There must be something noteworthy about a source called Futurism.com having the most consistently negative AI articles of anyone.  Here are two more.  The first is “There’s a Small Problem with the AI Industry:  It’s Making Absolutely No Money” (July 4^th).  Author Sharon Adarlo noted that “Goldman Sachs analysts have concluded… that AI just isn’t making any serious money yet,” and that “Goldman found that companies that hoped to profit from using AI to boost productivity – ranging from H&R Block to Walmart – have seen their shares vastly underperform the broader stock market since the tail end of 2022.”  She also found that “the only companies making much actual revenue off AI are the ones selling the hardware it needs, like Nvidia.”  While “companies using it want to see major returns,” “so far, it sounds like they aren’t.  Maybe the real question is how much runway the AI industry has before business leaders move onto the next thing.”  This piece is consistent with my previous comments.

The second Futurism.com piece was “Expert Warns That AI Industry Due for Huge Collapse” (Victor Tangermann, July 9^th).  He named a “founding partner” at a “macroeconomic research firm” calling AI “completely unproven,” saying that its hallucinations may never go away and that it was “too energy hungry,” along with a previous Stability AI CEO telling bankers that “this will be the biggest bubble of all time.”  Hardly universal views, but with merit.

Overall, is it true that, per New York Times columnist Thomas Friedman on July 9^th, “the artificial intelligence revolution of the past four years is widely expected to slam into the white-collar job market in the next four like a Category 5 hurricane”?  We don’t know that.  As with actual storms, we will benefit from tracking them and projecting their courses and times of landfall, but this one is way too weak and distant to fear.  We don’t need to buy flashlights, batteries, and water, and may never need to.  The future of artificial intelligence is still up in the air – and it may never come down.

Jobs in June: Seasonal Worsenings Mostly Offset Elsewhere, with AJSN Showing Latent Demand Up Almost Half a Million

The published projections I saw for this morning's Bureau of Labor Statistics Employment Situation Summary was that it would be worse than for May’s data a month ago.  The two estimates of net new nonfarm payroll positions were 190,000 and 200,000, and unemployment might be going up again.  So what happened?

Employment as above, at plus 206,000, was quite close to the predictions.  Seasonally adjusted joblessness had its third straight 0.1% gain, to 4.1%, with the corresponding total of people up 200,000 to 6.8 million.  (We now can ignore when the BLS says something “changed little.”)  Long-term unemployed, for 27 weeks or longer, gained 100,000 to 1.5 million, up 36% from June 2023, with so many people joining the labor force that its participation rate increased, 0.1% to 62.6%.  The measure of how many Americans are actually working, the employment-population ratio, stayed at 60.1%.  The count of those working part-time for economic reasons, or holding onto shorter-hours positions while looking for full-time ones, shed 200,000 to get to 4.2 million.  Average private nonfarm payroll earnings rose 9 cents per hour, close to the inflation rate, to $35.00.

Since May and June have different employment characteristics, the seasonally unadjusted figures did not match the others.  Unemployment that way jumped 0.6% to 4.3%.  The count of those not interested in working lost 637,000 to 93,776,000.  Those employed rose 433,000 to 161,774,000.

The American Job Shortage Number or AJSN, the statistic showing how many additional positions could be quickly filled if all knew they would be easy and routine to get, was up 478,000, as follows:

The share of the AJSN from those officially jobless was 4.3% higher at 37.7%.  Compared with a year earlier, the AJSN grew 504,000, with almost 800,000 more from unemployment partially equalized by, among others, 174,000 from fewer people wanting work but not looking for it for a year or more, and 200,000 fewer from expatriates. 

What patterns can we get from this report?  The new jobs, once again plentiful and nothing to take for granted, went largely to people with statuses other than simple unemployment.  Many more people returned to the labor market, and enough were unsuccessful to bring overall joblessness up.  A goodly number of those without work are not finding it, even after six months away.  Latent demand is not only alive and well but increasing.  Still, June is a tougher month than May, and the smaller, marginal categories show that this was, overall, a good one.  The turtle took a moderate step forward.