Sort:  
There are 4 pages
Pages

Three factors drive the advance of AI: algorithmic innovation, data, and the amount of
compute available for training. Algorithmic progress has traditionally been more difficult
to quantify than compute and data. In this work, we argue that algorithmic progress has
an aspect that is both straightforward to measure and interesting: reductions over time
in the compute needed to reach past capabilities. We show that the number of floatingpoint operations required to train a classifier to AlexNet-level performance on ImageNet
has decreased by a factor of 44x between 2012 and 2019. This corresponds to algorithmic
efficiency doubling every 16 months over a period of 7 years. Notably, this outpaces theoriginal Moore’s law rate of improvement in hardware efficiency (11x over this period). We observe that hardware and algorithmic efficiency gains multiply and can be on a similar scale over meaningful horizons, which suggests that a good model of AI progress should
integrate measures from both.

Have you ever considered that business may not have been that different 2000 years ago? Oh well, they did not have computers, internet and the digital economy but they still had a buoyant property market, they were trading across the seas and the land, and they were pretty good at financial administration. After all, empires were not built on air but on solid armies, bureaucracies and land expansion exercises (imperialism).

In the Roman World, the wealthiest citizens were focusing on the development of farming businesses. Several handbooks were written during the Republic and the Empire. These Agronomists followed the rich tradition of writers coming from the Greek times, such as Hesiod, Xenophon, Democritus of Abdera, even Aristotle. Even though the Roman writers followed the Greeks, they still managed to become better known and more widely revered than they predecessors. Their instructions became the ‘go to’ manual for Landowners across Europe for centuries to come. The most important works that survived until today are those of Cato, Columella, Varro, Virgil, Pliny and Palladius.

Today we will focus on Cato the Elder’s farming manual, De Agricultura. This is – not surprisingly – the eldest work of latin prose. It’s scope and importance emphasises on what actually mattered for the elite at the time and the manuscript dates from the second century BC, long before the Roman Empire reached the height of its power. In fact, long before Roman Emperors were ‘a thing’!

Let us take a look on what Cato believed of Farming as business. Obviously he preferred it over Commerce and Banking, both of which may have brought profits. However, the first was risky and the second was considered akin to usury and had the potential to confer upon the banker the title of ‘criminal’. As a result, the only moral and sensible option for the Roman Rich were to become farmers.

So, how did they go about it? First and foremost they needed to buy vast expanses of land. Or, alternatively, acquire them through wars. The Romans spent several centuries expanding their lands through military advances. So, there was a lot to be had. Secondly, they needed to learn how to manage these vast expansions of land. And this is where the Agronomists entered the picture. Cato was willing to present his contemporaries with a wealth of advice on how to set up and run a farming business. I will include a few of these comments just to give you an idea of what was expected of a Roman farmer.

As a rule, you should not be an Absentee Landowner. As the Master of the Household you should visit your farm often and upon arrival greet Lar (the Guardian/God/Ancestor of the Hearth). Always partner up with a God, just to be on the safe side. Once you pay your dues to Lar, it would be the right time to go around your property ON THE SAME DAY. Cato insists on this piece of advice and I suspect that the master would have liked to see how the farm is run on an average day and not after anticipating his visit. The element of surprise seems to be essential.

Once the initial checks have been completed, the next day can be dedicated to analysing the statistics “how much of the work is finished, how much remains, whether what is done was done in time and there will be time to do the rest, and how it is with the wine, the grain and everything else singly.” It should not astonish us that analytics were used regularly in antiquity in order to run a business. The owner needed to know how far they are in the production cycle and when the crops would be ready for sale.

Meme coins were some of the biggest winners after the U.S. presidential election, with some traders seeing it as a green light for a new crypto craze. Others have become worried that the latest Trump fueled meme mania was becoming too hot, however, and was likely to result not just in pain for investors but misallocation to less valuable projects in the industry.

Bitcoin losses Monday were relatively modest compared to meme coins and other smaller cryptocurrencies further out on the risk curve. It was last lower by just 3%, though it could see more pain in the short term as the trade war triggered by Trump's tariffs plays out.

GoCardless grew revenue by 41% to £132 million in full-year 2024. Of that total, £91.9 million came from customer revenue.

Last year also saw GoCardless record its first-ever month in profit in March 2024. Takeuchi said its his aim for GoCardless to post its first full-year profit in 12 to 18 months' time, adding it's "well on track" to do so.

The financial aspect would have taken even more time, because of its essential nature. After all, how else would they have known whether the business was profitable or even viable?And how would they have funded their elaborate lifestyles? Cato says on the matter: “You must check the figures for money and grain, check what is set aside for fodder, check the wine and oil figures — what is already sold, and the income from this, what is still to be produced, and what it will fetch — agree the difference and take charge of the agreed sum.” This is the end point (scope) of any business. Obviously, they did not have any double entry books (this was a much later invention) and they were lacking in cash flow projections. But they had an excellent understanding of basic finances and how numbers should be used for their advantage.

Back in September, GoCardless acquired a firm called Nuapay, which helps businesses collect and send payments via bank transfer.

Asked whether GoCardless is considering further mergers and acquisitions in future, Takeuchi said the firm is "actively looking," adding: "We're seeing lots of opportunities come up."

Following its acquisition of Nuapay, Takeuchi said GoCardless is currently testing a new feature that allows clients to distribute funds to their own customers.

"If you take something like energy, the vast majority of the payments are about collecting money," he told CNBC.

"But then you might have some of your customers that have solar panels on their roof and they're sending energy back to the grid, and they need to get paid for that energy that they're generating."

GoCardless, which is backed by Alphabet's venture arm GV, Accel and BlackRock, was last privately valued by investors at $2.1 billion in February 2022.

Takeuchi said the firm had no need for external capital and that there are "no plans" for an initial public offering in the near term.

As with every other business, delegation is a big part of running large organisations, including farms. The landowner probably had several farms that needed his attention. For each farm, he would have hired a manager to oversee the work on a daily basis. Even though the manager should have been a trusted employee (rarely a slave), this does not mean that he could waive his carte blanche and do whatever he wanted. Cato seemed to be very suspicious of the managers and insists on holding them accountable.

Fintechs have been watching Swedish fintech Klarna's plan to go public closely — but many are waiting to see how it goes before deciding on their own plans.

With technology IPOs at historic lows, several startups have instead opted to provide employees and early shareholders liquidity by selling shares in the secondary market.

In November, Bloomberg reported that GoCardless had chosen investment bank Lazard to advise it on a $200 million secondary share sale. GoCardless declined to comment on the report.

He eloquently writes on the subject: “ When you have this straight, you can get down to calculating people and days’ work. If the work seems wanting the manager will say that he has done his best, slaves were sick, the weather was bad, slaves ran away or were requisitioned for public works: when he has put these and all his other arguments, bring him back to the calculation of workers and their work! If there was rainy weather, what work could have been done while it rained? — washing and pitching vats, cleaning farm buildings, shifting grain, shovelling dung, making a dung-heap, threshing grain, mending ropes and making new ones; the slaves could have been patching their own cloaks and hoods.

On holidays they should have cleaned out blocked ditches, mended the public road, cut back hedges, dug the vegetable garden, cleared the meadow, cut sticks, pulled out brambles, husked the emmer, tidied up. While slaves were ill they ought not to have been given as much food.” It is astonishing the amount of detail he goes into. Just because the masters were wealthy, it did not mean they should not have intimate knowledge of shovelling dung, making heaps of them, ordering the planting of the vegetable garden, or cutting sticks. Quite the opposite! Even the food portions for the slaves were important, if profit was to be had!

Once the details were understood, then it was important to put forward the right orders for buying equipment, for selling the products (vegetables, animals or slaves), and for contracting the workers. The orders should have been delivered both verbally as well as in writing, so that there was a clear chronological record of the decision making process. Back then, papyri would have been the means of accounts. Today its a laptop with elaborate software. In both cases the outcome would have been similar.

The size of the penalties will depend on the infringement and size of the company fined.

That's higher than the fines possible under the GDPR, Europe's strict digital privacy law. Companies face fines of up to 20 million euros or 4% of annual global turnover for GDPR breaches.

The distillation of Catos’s wisdom (and probably his best advice) can be summarised in the following sentence “The master has to be a selling man, not a buying man.” This is the traditional advice of frugality and simplicity, where loans remained an anathema. In Cato’s advice to the manager (not the owner this time) he insists that: “He must lend to no one but ensure that the owner’s loans are repaid. He must have no loans out to anyone, of seed for sowing, food, wheat, wine or oil: there should be two or three households from whom he can ask necessities and to whom he can give, but no others. He must regularly make up accounts with the owner.” Although he does not moralise on the ‘evil’ of loans, he would not willingly accept such an agreement inflicted upon him, unless, of course, it is the owner that provides the loan

There’s widespread agreement there’s been impressive progress in AI/ML in the domains of vision, natural
language, and game playing in the last decade [Krizhevsky et al., 2012, Xie et al., 2016, Silver et al., 2018].
However, there’s massive disagreement as to how much progress in capabilities we should expect in the near
and long term [Grace et al., 2017]. For this reason, we believe measuring overall progress in AI/ML is a
crucial question, because it can ground the discussion in evidence. Measuring AI progress is critical to policymakers, economists, industry leaders, potential researchers, and others trying to navigate this disagreement
and decide how much money and attention to invest in AI.

It's worth stressing that the AI Act still isn't in full force — this is just the first step in a series of many upcoming developments.

Tasos Stampelos, head of EU public policy and government relations at Mozilla, told CNBC previously that while it's "not perfect," the EU's AI Act is "very much needed."

"It's quite important to recognize that the AI Act is predominantly a product safety legislation," Stampelos said in a CNBC-moderated panel in November.

"With product safety rules, the moment you have it in place, it's not a done deal. There are a lot of things coming and following after the adoption of an act," he said.

"Right now, compliance will depend on how standards, guidelines, secondary legislation or derivative instruments that follow the AI Act, that will actually stipulate what compliance looks like," Stampelos added.

In December, the EU AI Office, a newly created body regulating the use of models in accordance with the AI Act, published a second-draft code of practice for general-purpose AI (GPAI) models, which refers to systems like OpenAI's GPT family of large language models, or LLMs.

The second draft contained exemptions for providers of certain open-source AI models while including the requirement for developers of "systemic" GPAI models to undergo rigorous risk assessments.

All in all, the Romans did not put together elaborate business models based on debt and governmental grants. The clarity and minimalism of the economic and financial models at the time cannot be denied. Without trying to moralise on the topic, I would like to emphasise on the effectiveness of such an attitude. Today we are used to building large organisations based on loans, shares, use of derivatives. They grow fast and exponentially, they employ thousands of people and they go bust at the blink of an eye. We may enjoy the boosts as well as the busts of capitalism but I still appreciate the wisdom coming from the pre-industrial, pre-capitalist societies. Just for today, I intend to busk at the words of Cato and apply a bit more simplicity in my life!

Several technology executives and investors are unhappy with some of the more burdensome aspects of the AI Act and worry it might strangle innovation.

In June 2024, Prince Constantijn of the Netherlands told CNBC in an interview that he's "really concerned" about Europe's focus on regulating AI.

"Our ambition seems to be limited to being good regulators," Constantijn said. "It's good to have guardrails. We want to bring clarity to the market, predictability and all that. But it's very hard to do that in such a fast-moving space."

Still, some think that having clear rules for AI could give Europe leadership advantage.

Most people who read ancient history are familiar with the kings of Rome, but the republic and empire get the lion’s share of the attention and the kings are usually relegated to mere anecdotes. Still, it’s interesting to discuss the kings and their part in founding the republic.

Why did Rome have kings, and why did they lose power in favor of the republic? We’ll discuss these questions later, but only peripherally, because the goal of this article is to talk about the kings and what we really know about them.

The Roman kings are shrouded in myth and invention: myth because Rome wanted to create a mythology like the Greeks; invention because much of the history was embellished to support the image of Rome. Here we will discuss what is known and toss the rest.

"While the U.S. and China compete to build the biggest AI models, Europe is showing leadership in building the most trustworthy ones," Diyan Bogdanov, director of engineering intelligence and growth at Bulgarian fintech firm Payhawk, said via email.

"The EU AI Act's requirements around bias detection, regular risk assessments, and human oversight aren't limiting innovation — they're defining what good looks like," he added.

Our main source for information about the Roman kings is Livy Book I-V, the History of Rome. Livy’s dates were 59 BCE – 17CE, so he was writing about events up to 700 years before his time.

Livy was the first professional historian in Rome; historians who preceded him were wealthy people who studied history as a recreational activity. Livy read Thucydides and adopted his methods of relating history as stories about people. Those stories were made interesting to hold the attention of the reader. Livy believed Thucydides’ theory that history was a series of repeatable events displaying human behavior. For example, a tyrant in 500 BCE would act the same as a tyrant in 100 BCE. That idea made it easy to transfer current facts into the past. If the current tyrant stole money from the treasury, then the tyrant, from 400 years before, must have stolen money from the treasury.

Bitcoin's down move was more modest than that of other cryptocurrencies. Ether plunged 12% to around $2,600. It was trading above $3,300 Friday. Meme coins were among the hardest hit.

Jeff Park, Bitwise Asset Management's head of alpha strategies, said a sustained tariff war would be "amazing" for bitcoin in the long-run due to an eventual weakening of the dollar and U.S. rates.

There were certainly events that made an impression on the Roman people, who wrote them down and carried them through history, so Livy has some true facts to use. One example was the barbarian sack of Rome in 390 BCE. That story was well-known to all Romans. A treaty between Rome and Carthage dated 507 BCE was also verified, as was a solar eclipse in 404 BCE.

We know that Rome was originally settled by two separate groups who tended goat and cattle herds, because of the differences in their pottery. They lived in huts on the top of two of the famous Roman hills: the Palatine and the Esquiline and lived a pastoral life, which continued for a century or more until the Etruscans appeared. More on that later.

Let us review the history of the kings in chronological order. Those stories that are bolded appear to be true.

While many believe bitcoin is a hedge against inflation and uncertainty over the long term, it trades like a risk asset in the short term — and could endure further pain this month due to uncertainty around the trade war triggered by Trump's tariffs.

"Digital assets will eventually like today's [U.S. Treasury] yield mix (higher break-evens and lower real yields) but it will take outright nominal yields to roll lower at some stage (on growth fears) to solidify that," Geoff Kendrick, an analyst at Standard Chartered, said in a note Monday. "Until then we may be in for a choppy few days where the $90,000 level in BTC is again at risk."

  1. Romulus (753-716), the founder of Rome, did not exist, but he was an important character in Roman mythology. The story of Romulus and Remus, with different names, was a Greek legend. There is no factual evidence about the existence of Romulus and the traditional founding date of the city, 753 BCE, was arbitrary.

  2. Numa Pompilius (715-672) established Rome’s religious traditions, built temples, and set down rules for worshiping the gods. All Romans knew him as the king who created the cult of the Vestal Virgins. Numa introduced a legal system that governed some aspects of Roman life, such as marriage and contracts. The story of the Vestal Virgin story is the only plausible one.

There are 4 pages
Pages