Sunday, July 28, 2019

Siri, Cortana, Alexa and AI (Weekend Humor)




Not there yet but will our online assistants be the first to show early signs of intelligence? That would be quite amazing because such AI would be stumbled into instead of created by design. This is possible but unlikely. There must be a few basic principles such as backward propagation which still elude us. Then there is pure intelligence where the right deductions are generated without awareness and a more advance system with a self which "thinks" nonstop without being stimulated. As we progress, the steps will probably be more numerous and complex. Jealousy and humor are still our prerogative for some time! 

Friday, July 26, 2019

The death of third party data







Interesting article published on Technative
https://www.technative.io/the-death-of-third-party-data/

Interesting but incomplete!
There is a mix up of understanding between 3rd party using the data for access to clients and companies using 3rd party data to get more insight into the clients. This is not the same!

In one case, you have intrusion and this is what people react to quite legitimately. In the other case, we are talking about necessary insight into people behavior in order to do marketing properly. Without insight and context, you cannot do "personalized" marketing.

Marketing personalization will more and more rely on AI systems and these systems will need more and more data to perform efficiently. Complex models may eventually replace data but these have to be built and to do that you will need data, lots of data!



With GDPR in full force, the days of third party data are numbered

The regulation has been the catalyst for new concerns around how companies are using our personal information, and there is heightened awareness around how anonymous third-party data cookies are tracking us around the internet. In a post-GDPR world, sensitivity to intrusive online ads has never been so strong.
And it’s a justified cause for concern. Imagine going into H&M and someone walking up to you to sell a Primark T-shirt. It wouldn’t feel right, but that’s exactly what ad space is doing. You could be on a travel website and another completely different company, with no affiliation, not only knows that you’ve been there but that you’ve made a purchase.
This abuse of third party data has become the norm, with too many companies crossing the line and violating consumer’s privacy. And let’s not forget, this practice has been technically possible and legally allowed. What Facebook and Cambridge Analytica did, for example, didn’t break the law. They abused something they were allowed to abuse. But consumers have become savvier when it comes to their data and they rightly called ethics into question.
In this climate, marketers need to be prepared for a backlash on a much wider scale, and for the spotlight on ethics to kill acquisition marketing. Data Management Platforms (DMPs), which currently allow marketers to access huge volumes of third-party data way beyond the resources of the individual marketer, will become a thing of the past. Or marketers will at least have to accept that there will be radical changes or limitations for these platforms to function within GDPR regulations.

Putting retention marketing centre stage

This will put significant pressure on CMOs to shift their focus to providing the best possible experience to existing customers, targeting known email addresses and mobile numbers instead of using cookie crumbs. The beauty of retention marketing is that the brand has already won the customer over at some point in the past, so they are likely to be more receptive to personalised content and engagements.
Forrester research shows that the majority are already onto this trend. In 2017, CMO spend on customer growth and retention outpaced budgets for customer acquisition by an almost 2:1 ratio (63% and 37% share of budgets, respectively). That’s good progress when we consider that ten years ago the use of third party data dominated. But there’s still a long way to go and marketers should be working to get their acquisition budgets right down to meet customer expectations.

Turning first party data into a gold mine

The next challenge for marketers is to maximise spend on retention marketing and make first party data a powerful addition to their omnichannel strategy. For example, as customers voluntarily follow a brand’s social channel, it makes sense to tap into this as much as possible. Retention marketers have the opportunity here to transform communication and capitalise on tighter customer relationships across the digital journey.
In addition, CRM-based advertising enables marketers to use their first-party contact data to reach anyone online, wherever they are, with relevant ads. Marketers can extend their reach across networks without diluting targeting focus, and can drive engagement to provide comprehensive, exciting customer journeys. The critical aspect here becomes the marketer’s ability to accurately match first-party data with network profiles to target users on networks such as Google and Facebook.
However, what will ultimately drive revenue is automating the integration of first-party data. Marketing automation needs to take a strategic approach in order to be successful, as poorly implemented automation can very quickly ruin personalisation.
In a post GDPR world where consumers expect – and deserve – a more responsible use of their data, the key to success is building personalised customer journeys that inspire customers to keep engaging with the brand. By combining first-party data with sophisticated marketing platforms, and being aware of the changing digital habits of consumers, marketers now have all the ingredients they need to seriously transform their online strategy and win at retention marketing.

Monday, July 22, 2019

12 Free Machine Learning Course

This is an amazing collection of tools about Machine learning, so although this post is a little long, I list it in its entirety as a repository, including all the links to the different resources.

The Original can be found at:

on Techgrabyte at https://techgrabyte.com/





12 Machine Learning Course For Free That Will Make You An Expert


Machine Learning is a beautiful field to work on, it is full of fun and if you are one who is looking to learn ML, then you are at right place, here today I’ll show some of the best machine learning course that will not only save your money but it will also offer you a quality education.
And the fun part is that you don’t even have to leave your room, just sit down on your chain, pull the desk closer, take a notebook and pen and finally grab the most import thing, a cup of coffee and go full-on it.


This all machine learning courses are perfectly designed and divided into a curriculum that will take you from an absolute beginner to an expert in Machine Learning, who knows how to build real-world projects.
From the listed machine learning course, you will learn the basics and fundamentals of Machine Learning, how it works internally, how to train a model and also how to implement the knowledge that you will gain.

For a general idea, this machine learning courses will teach you about various different concepts like parameter learning, logistic regression model, neural networks, application of neural networks, cost function, and backpropagation and much more.
Another thing is that all the machine learning course that we are going to see here are primarily selected on the quality of content, language simplicity, and public reviews.
The only prerequisites that are required for this all machine learning courses are the little understanding and knowledge of programming languages like R and Python and little basics about mathematics.
That been said, let us start our journey. Don’t forget to check the bonus.



1. Programming Language

12 Machine Learning Course For Free That Will Make You An Expert
If you have already master or learning a programming language that heavily going to contribute in creating Machine Learning based stuff, then you can directly move to the next point, but if you are new here then wait I have a surprise for you.
A blank mind is always good as it is free from confusion, if you haven’t learned any language yet, then at this right moment start with Python or with R or with JavaScript, I’ll suggest you Python.
I mostly suggest beginners go with Python for Machine Learning, the prime reason behind this is that Python is quick, fast and easy to understand and learn.
Medium has a great article on Why Python is the most popular language used for Machine Learning, I’m so mad for Python that I can talk about it all day, but to keep the article short you can head over this article and can clear doubts and questions, you can also ask them in below comment box.
And if I talk about R, then R is good for ad-hoc analysis and exploring data sets, R has a steep learning curve, but people without programming experience, find it overwhelming.
R also don’t have lots of libraries that Python actually offers, also the community of Python is humongous, so you can easily find someone who can help you in your errors.
Now the surprise that I talk before.
If you are thinking to learn R or Python, then below are the links from where you can download the most popular and useful books for Python and R for free.
This book will help you to build the core foundation on which you can learn advanced concepts at a greater speed.

2. Udacity’s Intro to Machine Learning

The clarity in thoughts is must when you are learning something, especially when it is Machine Learning. Udacity’s Intro to Machine Learning is the best machine learning course that will teach and will show you what machine learning actually is.
Another reason that I suggest this machine learning course is that I many time saw that the students have their some pre-assumptions, confusion, and doubts on machine learning, Udacity’s this course will give you a better understanding of ML.
This course is about 10-weeks long and it will teach you the end-to-end process of investigating data through a machine learning lens.
It will teach you how to extract and identify useful features that best represent your data, the most important machine learning algorithms, and how to evaluate the performance of your machine learning algorithms.
I also recommend you to take the foundational Intro to Data Science course which deals with Data Manipulation, Data Analysis, Data Communication with Information Visualization, and Data at Scale, this will helps you understand machine learning concepts more easily.
Intro to Machine Learning will be taught by instructors Sebastian Thrun and Katie Malone, the instructors expect the beginners to know basic statistical concepts and Python.




3. Stanford Machine Learning Course

Stanford Machine Learning Course
This machine learning course is a very special course, as it is taught by Andrew Y. Ng (my idol), Andrew is a prestige name in the field of machine learning. He is a co-founder of CourseraBaidu’s Chief Scientist and a former head of Google Brain.
In this machine learning course, you will learn about the most effective machine learning techniques.
You will learn about some of Silicon Valley’s best practices in innovation as it pertains to ML and AI. This course provides a broad introduction to machine learning, data mining, and statistical pattern recognition.
Topics that these machine learning course will cover include:
(i) Supervised learning, parametric and non-parametric algorithms, support vector machines, neural networks.
(ii) Unsupervised learning clustering, dimensionality reduction, recommender systems, deep learning.
(iii) Best practices in machine learning -bias/variance theory, innovation process in machine learning and AI.


This machine learning course also offers numerous case studies and applications so that you’ll learn how to use ML algorithms to building smart robots (perception, control), text understanding (web search, anti-spam), computer vision, medical informatics, audio, database mining, and other areas.
After doing this course, you will gain skills like Logistic Regression, Artificial Neural Network, and Machine Learning.

4. Google Machine Learning Course

Google Machine Learning Course
With no doubt, Google is the leading company in the field of ML, and Google is offering some amazing machine learning courses that is absolutely free and that will teach you many valuable concepts of ML.
This free Machine Learning Course will teach you how to recognize the relative impact of data quality and size to algorithms, set informed and realistic expectations for the time to transform the data.
This course will also explain you a typical process for data collection and transformation within the overall ML workflow, how to collect raw data and how to construct a data set, and split your dataset with considerations for imbalanced data and about transform numerical and categorical data.
This course will have 25 lessons, 40+ exercises, interactive visualizations of algorithms in action, real-world case studies, and mind-blowing lectures from Google researchers.
This course will answer all questions like :
1) How does machine learning differ from traditional programming?
2) What is a loss, and how do I measure it?
3) How does gradient descent work?
4) How do I determine whether my model is effective?
5) How do I represent my data so that a program can learn from it?
6) How do I build a deep neural network?

5. Mathematics for Machine Learning Specialization

Mathematics for Machine Learning Specialization
Mathematics is the foundation of ML, all the algorithms and programs that you will write, will somehow always be directly or indirectly related to Mathematics.
It is the foundation of a student learning ML, and this course called “Mathematics for Machine Learning Specialization” is all designed to make your foundation solid strong.
It is one of the most recommend machine learning course by experts. This course will help in getting speed in the underlying mathematics, building an intuitive understanding, and relating it to Machine Learning and Data Science.
This course will teach you about what is linear algebra and how to it relates to data, what are vectors and matrices and how to work with them and use them for data fitting.
In the second section called Multivariate Calculus of this course, you will learn about how to optimize fitting functions to get good fits to data, what is calculus and how to uses it.
The third section, Dimensionality Reduction with Principal Component Analysis, will teach you about how to use the mathematics from the first and second sections, and how to compress high-dimensional data.
This course is of intermediate difficulty and will require basic Python and Numpy, knowledge, which you can quickly learn through above-mentioned books or from youtube.
At the end of this specialization, you will have gained the prerequisite mathematical knowledge to continue your journey and take more advanced courses in machine learning.

6. Machine Learning Fundamentals

Machine Learning Fundamentals
Machine Learning Fundamentals is a course offered by the University of California, San Diego.
In this 6th machine learning course, you will learn many different things and you will also use the knowledge that you gain from previous courses.
This course will be taught by Sanjoy Dasgupta, Professor of Computer Science and Engineering, UC San Diego.
In this course, you will learn about the classification, regression, and conditional probability estimation, generative and discriminative models, linear models and extensions to nonlinearity using kernel methods.
You will also learn about ensemble methods, like boosting, bagging, random forests, representation learning: clustering, dimensionality reduction, autoencoders, deep nets.
The reason behind the selection of this course is that it also offers real-world case studies of Machine Learning.
By this real-world case studies you will learn how to classify images, identify salient topics in a corpus of documents, partition people according to personality profiles, and automatically capture the semantic structure of words and use it to categorize documents.
This course also offers an additional section called Data Science MicroMasters program, where you will learn a variety of supervised and unsupervised learning algorithms and the theory behind those algorithms.
Note that all programming examples and assignments will be in Python, using Jupyter notebooks.
After successfully learning this course, you will be able to analyze many different types of data and to build descriptive and predictive models.

7. Machine Learning: Classification

Machine Learning: Classification courses



Machine Learning: Classification is one of the most enjoyable and knowledge riched machine learning course on this list.
It is divided into two sections, the first one is based on analyzing sentiment, where you will create models that predict a class (positive/negative sentiment) from input features (text of the reviews, user profile information).
The second section will have loan default prediction, where you will tackle financial data, and predict when a loan is likely to be risky or safe for the bank.
These tasks are examples of classification, one of the most widely used areas of machine learning, with a broad array of applications, including ad targeting, spam detection, medical diagnosis, and image classification.
This machine learning course will teach you how to create classifiers that provide state-of-the-art performance on a variety of tasks.
You will become familiar with the most successful techniques, which are most widely used in practice, including logistic regression, decision trees and boosting.
The major things this Machine Learning will cover be the input and output of a classification model, tackle both binary and multi-class classification problems, implement a logistic regression model for large-scale classification.
It will also teach you how to create a non-linear model using decision trees, improve the performance of any model using boosting and how to scale your methods with stochastic gradient ascent.
In addition to this, you will be able to design and implement the underlying algorithms that can learn these models at scale, using stochastic gradient ascent.
You will implement these technologies on real-world, large-scale machine learning tasks.
You will also address significant tasks, and real-world applications of ML, including handling missing data and measuring precision and recall to evaluate a classifier.
This course is hands-on, action-packed, and full of visualizations and illustrations of how these techniques will behave on real data.
It also included optional content in every module, covering advanced topics for those who want to go even deeper.
The concepts in this course will be implemented with Python.

8. Machine Learning: Regression

Machine Learning: Regression machine learning courses
In this machine learning course, you will learn more about regression in ML, regression is a measure of the relation between the mean value of one variable and corresponding values of other variables.
This course is divided into a number of sections in order to build an easy understanding.
The first section will be of predicting house prices, where you will create models that predict a continuous value from input features like square footage, number of bedrooms and bathrooms.
Other applications range from predicting health outcomes in medicine, stock prices in finance, and power usage in high-performance computing, to analyzing which regulators are important for gene expression.
In this machine learning course, you will explore regularized linear regression models for the task of prediction and feature selection.
You will be able to handle very large sets of features and select between models of various complexity.
You will also analyze the impact of aspects of your data such as outliers on your selected models and predictions. To fit these models, you will implement optimization algorithms that scale to large datasets.
The major things this Machine Learning will cover will be the input and output of a regression model, compare and contrast bias and variance when modeling data, estimate model parameters using optimization algorithms.
It will also cover tuning parameters with cross-validation, analyze the performance of the model, describe the notion of sparsity and how LASSO leads to sparse solutions, build a regression model to predict prices using a housing dataset.
The concepts in this course will be implemented with Python.

9. Machine Learning for Data Science and Analytics

Machine Learning for Data Science and Analytics machine learning courses
Though machine learning is sub-branch of AI, I find it mostly close to Data Science.
When you learn machine learning and play with it you will find that you are mostly interacting with data and using all its concepts, and at some point, you will also start feeling this way.
These days machine learning is now mostly using for searching the web, placing ads, credit scoring, stock trading and for many other applications, by doing this course you will also be able to do such things.
This data science course is an introduction to machine learning and algorithms.
From this machine learning course, you will develop a basic understanding of the principles of machine learning and derive practical solutions using predictive analytics.
This course will also help you understand why algorithms play an essential role in Big Data analysis.
In this course, you will learn what machine learning is and how it is related to statistics and data analysis.
At the start of this course, you learn how machine learning uses computer algorithms to search for patterns in data, how to use data patterns to make decisions and predictions with real-world examples from healthcare involving genomics and preterm birth.
It will also teach you about how to uncover hidden themes in large collections of documents using topic modeling, how to prepare data, deal with missing data and create custom data analysis solutions for different industries, basic and frequently used algorithmic techniques like sorting, searching, greedy algorithms and dynamic programming
This course will be taught by Ansaf Salleb-Aouissi, Cliff Stein, David Blei, Itsik Peer, Mihalis Yannakakis and Peter Orbanz.

10. Foundations of Data Science: Prediction and Machine Learning

Foundations of Data Science: Prediction and Machine Learning courses
I know you might be getting sucked or bore with predictions and regression, but believe me, this machine learning course will definitely help you to level up your knowledge and skills that you have achieved through previous courses.
This course called “Foundations of Data Science: Prediction and Machine Learning” mainly focuses on regression and classification, to automatically identify patterns in your data and make better predictions.
One of the principal responsibilities of a data scientist is to make reliable predictions based on data. When the amount of data available is enormous, it helps if some of the analysis can be automated.
Machine learning is a way of identifying patterns in data and using them to automatically make predictions or decisions. In this data science course, you will learn the basic concepts and elements of machine learning.
The two main methods of machine learning you will focus on are regression and classification.
Regression is used when you seek to predict a numerical quantity. Classification is used when you try to predict a category e.g, given information about a financial transaction, predict whether it is fraudulent or legitimate.
For regression, you will learn how to measure the correlation between two variables and compute a best-fit line for making predictions when the underlying relationship is linear.
The course will also teach you how to quantify the uncertainty in your prediction using the bootstrap method. These techniques will be motivated by a wide range of examples.
For classification, you will learn the k-nearest neighbor classification algorithm, learn how to measure the effectiveness of your classifier, and apply it to real-world tasks including medical diagnoses and predicting genres of movies.
The course will highlight the assumptions underlying the techniques and will provide ways to assess whether those assumptions are good. It will also point out pitfalls that lead to overly optimistic or inaccurate predictions.
The major concepts that this course covers are fundamental concepts of machine learning. linear regression, correlation, and the phenomenon of regression to the mean, classification using the k-nearest neighbors’ algorithm, how to compare and evaluate the accuracy of machine learning models, basic probability and Bayes’ theorem.
This course will be taught by Ani Adhikari Teaching Professor of Statistics UC Berkeley,  John DeNero Giancarlo Teaching Fellow in the EECS Department UC  Berkeley, David Wagner Professor of Computer Science UC Berkeley.
Enroll Today

11. Dynamic Programming: Applications In Machine Learning and Genomics

Dynamic Programming: Applications In Machine Learning and Genomics
If you look at two genes that serve the same purpose in two different species, how can you rigorously compare these genes in order to see how they have evolved away from each other? And the answer to this question will be in this 11th machine learning course.
Our 11th machine learning course will be taught Pavel Pevzner Ronald R. Taylor Professor of the Computer Science the University of California, San DiegoPhillip Compeau Assistant Teaching Professor Carnegie Mellon University.
In this course, you will get introduced to Algorithms and Data Structures MicroMasters program, where you will see how the dynamic programming paradigm can be used to solve a variety of different questions related to pairwise and multiple string comparison.
This course will teach you dynamic programming and how it applies to basic string comparison algorithms, sequence alignment, including how to generalize dynamic programming algorithms to handle different cases, hidden Markov models.
It will also teach you about how to find the most likely sequence of events given a collection of outcomes and limited information, Machine learning in the sequence alignment.
In the second part of the course, you will see how a powerful machine learning approach, using a Hidden Markov Model, can dig deeper and find relationships between less obviously related sequences, such as areas of the rapidly mutating HIV genome.
This is one of the amazing course we have up to in this list of best machine learning courses.

Big Data Keeps Getting Bigger



The sun never sets on the creation of new data.
Authored by Jeff Desjardins via VisualCapitalist.com,

How can we grasp and understand big data?
This often seems to be a challenge especially for people who do not understand it... and nevertheless try to explain it to us!

This is the case here with a hodgepodge of disparate data sources and data usage packaged together in a colorful but not quite meaningful way to be impressive, leading to the wrong conclusion:

"Also, imagine how difficult it is to make sense of this swath of data. How does one even process insights from the many billions of Youtube videos watched per day?"

Well, no actually! With big data, it is the exact opposite: The more data you have, the easier it is to make sense of it!

And this is in fact relatively easy to understand: The more data you have, the easier it becomes to find patterns, trends and correlations. 

Many companies struggle with their big data projects, often because the data is not big enough, and most often because it does not contains important clues or information preventing the right insight to be accessed. 

So as the data flow increases, the difficulty is far more technical in a hardware  kind of way than in the understanding of the data as implied in the article above. 

But then, once you understand the contents comes of course the real challenge of our time: What to do with the data. This, more than being mesmerized by peta and exabytes is or should be the focus of our attention.  

Saturday, July 20, 2019

'Stalling Markets': The Last Time This Happened Was October 1929


'Stalling Markets': The Last Time This Happened Was October 1929
Fri, 07/19/2019
Great article from GoldMoney, focusing on the US and the mercantilist policies of the Trump administration.
It is unfortunate that China is not included in the broad picture as the Chinese economy is also slowing down fast with a highly leveraged financial system.
"History does not repeat, it rhymes!"
October 1929 it may not be but there can be little doubts that we are approaching a new paradigm.





The combination of American trade protectionism and the end of a failing credit expansion is leading into a global economic downturn, and potentially a systemic crisis. Meanwhile, investors still believe more extreme monetary policies will stabilise economies and that the ultra-low interest rate environment will persist without renewed price inflation. As Samuel Johnson reputedly said of a second marriage, it represents the triumph of hope over experience.

Introduction
There is a moment just after the top of every credit cycle where positive momentum stalls before a new reality emerges. When the stall begins, as appears to be the case today, everything is still read positively. Perennial bulls say “Don’t worry, the central bank will reduce interest rates and inject enough money into the banking system to ensure any recession will be minor and growth will resume”. With interest rates falling, confidence in the final outcome means stocks continue to rise. With this mindset, bad news for the economy is always good news for stocks.
This investors’ paradise is populated by devotees of the new economics, supporting progressively increased state intervention. They don’t actually believe that free markets should set stock prices anymore and have become hooked on central banks pursuing inflationary policies. In their minds, the relationship between monetary inflation and rising stock prices amounts to a financial equivalent of perpetual motion. However, their enduring belief in the might of central banks and the importance they place on maintaining asset prices makes inflationists blind to the message from stalling markets.
We all get caught up in it. And when evidence of the stall in economic growth mounts, we clamour for lower interest rates, credit expansion, and finally competitive exchange rates. Even President Trump is now telling us the Fed must weaken the dollar to boost exports and the American economy. As it is already doing his bidding with interest rates, surely the Fed will oblige.
The naivety of this reasoning is endemic, and as a naïve supporter of free markets, President Trump is beginning to trot it out again. In Britain, the same old inflationist story, wrapped up in proposed tax cuts to be paid for later by economic growth, is now being pushed by Boris Johnson, almost certain to be the next Prime Minister. Let us hope this is just electioneering rhetoric. But so widely are the myths of monetary stimulus believed that they are now certain to be renewed in a push to sustain economic growth.
That gold prices measured in dollars will be guaranteed to rise is indirectly becoming officially sanctioned. Those of us in the gold business would be grateful for President Trump’s endorsement of gold if it was not for the economic consequences of his weak dollar policies on ordinary people whose money is about to be trashed. Nevertheless, the gold price has now jumped to a new level and markets are trying to discount the consequences of what is now unfolding. As the gold price continues to rise, those who wonder whether it is worth buying miss the point. It is not gold that rises, but their money that’s falling. Money is falling because governments through more aggressive monetary policies are about to deliberately undermine their own currencies.
The naivety over the consequences of weak money is not restricted to the leaderships of the US and UK. The maintenance of negative interest rates and bond yields in the EU and Japan are already testament to that. But we now see two world leaders (assuming Boris succeeds May) who indicate they are aware of the failings of the economic establishment, but publicly endorsing the inflationism that has been central to the establishment’s failure. By promising the tax jam of today for a better outlook tomorrow, Boris Johnson is roughly where Trump was before he was elected president. Let’s hope an enterprising journalist asks him to clarify whether he thinks a lower sterling (which we already have) is good for the economy. Very likely, he will agree, perhaps with the proviso that price inflation remains under control.
Therefore, central banks are committed to address stalling economic growth by snorting more of the drug that is giving us the downer. But our leaders and central bankers are ignorant of sound theories of money and credit and are driven almost entirely by statistical information. And here we hit a further problem. Government statistics are not fit for purpose.

Statistics are misleading
Statistically, inflation is under control, because the statisticians with their methods have ensured it is so. In giving the central banks a passport to accelerate the rate of monetary inflation by suppressing the consequences, they are storing up trouble for us all. There will come a time when not even the manipulation of consumer price statistics will hide the fact that the purchasing powers of the dollar, sterling, euro and even the yen are all falling at an alarming rate. It is when we begin to sense that this is a problem that the stalling feeling turns our greed, or complacency, into concern and then outright fear.
Exactly two hundred years ago, Lord Canning, who was briefly prime minster of Great Britain, warned that you can prove anything with statistics, except the truth. This was over a century before modern econometrics evolved to make statistics more meaningful. But statisticians have missed what was behind Canning’s point: statistics prove nothing because they cannot replace sound reasoning. Instead, they only assume the relationships between cause and effect. Rather than attempting to understand the manner of the link, their approach is to monitor the relationship between monetary inflation and changes in prices, to ensure that increases in the CPI remain within a target range. As long as this is the case, monetary policy makers can carry on issuing money out of thin air.
You cannot measure the general level of prices, because it is a concept and not a fact. Strictly speaking, it is hardly a concept either, because it is not the unmeasurable general level of prices that inflates, but the purchasing power of the currency that diminishes. It may take time to work through, but if you increase the quantity of unbacked state money and the bank credit in circulation, a unit of currency will simply buy less.
Not enough recognition is given to the draining effect on the productive members of society and their businesses. Monetary inflation undermines the value of their earnings and profits, transferring wealth from savers, and it impoverishes the poorest labourers for the benefit of government finances.
Far from obtaining something for nothing, the government gets its seigniorage by impoverishing the same people that pay its taxes. If one could measure the general level of prices, it is more likely that they have been rising by between seven and ten per cent annually for a considerable time, as illustrated by John Williams’s ShadowStats, and confirmed by the Chapwood Index. The official two per cent target is poppycock. If we assume the two independent calculations are more realistic, US citizens have been getting collectively poorer every year since the financial crisis of 2008/09. Not only is this evidence that Canning’s aphorism about statistics telling us everything but the truth still holds, but governments now fully depend on the concealment of the true state of affairs by statistical suppression.
Governments have arrived at this point because funding through inflation faces the law of diminishing returns. The more a government inflates, the more it impoverishes its people. And the more the people are impoverished, the less both taxes and inflationary financing yield. And as we look down from the heights of inflated asset prices, the more evidence emerges that our economies are stalling, the more important this will become.
With the establishment, investing institutions and regular investors all being misled by official statistics, it is no wonder that an understanding of the true position hardly exists. It is an Alice-in-Wonderland world where the more you inflate, the more GDP statistics say the economy is growing. Under-recording the price deflator has become central to maintaining the delusion. Almost no one realises that an increase in nominal GDP is no more than a reflection of more money and credit being injected into the economy. It confuses this increase in the quantity of money and credit with progress. But progress suffers the disadvantage, like the general price level, of being impossible to measure with statistics.
Reliance on statistical method has encouraged wrongheaded government intervention. Long ago, we dismissed the certain knowledge that society thrives by cooperation and governments only by intervention. The former progresses, the latter interferes. It prevented governments from trying to improve on free markets, but that ended following the First World War. Consequently, decades of intervention from the 1920s onwards have increasingly distorted our world away from free-markets to embrace the Gospel of Government. The gospel has been a drip-feed upon which modern economies have become increasingly dependent.

Stalling into a nose-dive
That drip-feed is now augmented by American trade protectionism, reversing the expansion of trade from which we have all benefited. The harmful effect on the American economy will become apparent. The combination of a long period of credit expansion and trade tariffs will very likely drive it into a deepening recession, possibly a slump, as these conditions today repeat those of 1929-1932.
If the inflationary effect on prices is to be limited, it will require foreign investors to buy dollars and increasing quantities of US Treasury debt to cover an escalating budget deficit. Global funds will have to be diverted to the dollar from other investment opportunities, notably the widescale development of Asia. From its policy towards China’s economic development, elements of the American deep state appear to understand this. President Trump appears not to. And now, he proposes to weaken the dollar the foreigners are expected to buy to finance his escalating budget deficits.
It is a difficult trap he has unwittingly set for his administration. And as the US economy stalls further, and the dollar weakens in a vain attempt by policy-makers under The Donald’s cosh to make America great again, the dollar’s slide will require rising interest rates for its purchasing power to be stabilised, forcing US Treasury prices into a bear market. The US Treasury’s finances will be plainly ensnared in a debt trap.
Other currencies, driven for decades by the same Keynesian logic, are to greater and lesser extents in the same boat. But every currency has two driving forces that determine their valuations. There is the collective assessment of the foreign exchanges, and changes in preferences between holding money and buying goods in the domestic economy. Sometimes, the foreigners might feel a fall is overdone, and buy a currency when its domestic purchasing power is still falling. At other times, the slide towards oblivion is deferred by the general public who cannot get their heads round what is happening to their government’s money. But despite these interacting forces, once the world’s reserve currency begins to decline, interest rates everywhere have to rise.
If this increasingly likely event happens, the effect on forward-looking markets is certain to be brutal. Today’s stall becomes a free-fall tomorrow. That is why it is likely that by the end of this year it will be increasingly apparent that national economies, emasculated by continual wealth-transfer through monetary debasement and over-burdened by non-productive debt, will begin to rhyme with the crash of 1929-32 and the subsequent depression. The most notable difference is that with today’s currencies being unbacked fiat instead of tied to gold, prices will rise instead of falling as they did in the 1930s.

Implications for gold
The last time the destructive forces of an end-of-credit-cycle coincided with trade protectionism was in October 1929. They were the driving factors behind the Wall Street crash and the subsequent depression. This time, the tariffs are not nearly as high as those of the Smoot-Hawley Tariff Act, but the magnitude of the credit cycle is far greater. While we can hope that this time the combination is not as disruptive as the 1929-32 episode, there is no doubt that today there is enough of a build-up of market distortions ready to wash out of the global economy to justify considerable unease.
This unease is yet to be manifest in widespread investment opinion, which still hopes for a miracle from monetary policies. But both our analysis and the empirical observations of events ninety years ago demonstrate why a miracle is impossible. A slump in global business activity is already developing, and the only policy response will be inflationary. Monetary expansion is effectively guaranteed in a vain attempt to stop a downturn and to ensure the banking system is preserved. This compares with a 25% contraction in broad money between 1929 and 1933 as thousands of American banks went under.
Through the medium of the dollar, in 1929-33 prices were measured in gold, which was fixed at $20.67 to the ounce. This time, there is no sheet anchor, and the dollar will simply lose purchasing power. This means there will be more dollars to the ounce of gold. There is no point in speculating how many dollars there will be to the ounce; you might as well debate how many angels can dance on the head of a pin. More importantly, it is difficult to see how the slide in the dollar’s purchasing power can be stopped once it starts.
Just as the ability of the productive sector to pay taxes is being increasingly undermined at the same time as the government’s expenditure rises, we can also see time preferences adding a further layer of destruction to government finances. Foreigners in particular will need far higher interest rates to stop them selling dollars and to persuade them to buy again. This was the policy of Paul Volcker, as Chairman of the Fed addressed in the early eighties, when he increased interest rates to 20%. Federal government debt then stood at only 30% of GDP, while today it stands at 105%. Putting aside the bad debts escalating at the banks from a grossly overindebted private sector, a rise in interest rates sufficient to stabilise the fiat dollar would almost certainly wipe out government finances and therefor faith in the dollar itself.
That is the extent the debt trap has now reached, and the problem is not confined to America. All major economies are in the same boat with very few notable exceptions. The Eurozone includes governments with severe debt problems, and the Japanese government has the highest debt to GDP ratio of them all. Lesser currencies have always had difficulties, which will simply escalate if dollar interest rates rise.
For the moment, very few see the true extent of the fiat currency problem. It is hard for them to visualise an economic slump when overall demand for goods drops, and for their prices to rise at the same time. They are fixated on the objective value of money in transactions, and do not realise that if people lose faith in it, a currency’s purchasing power will slide.
When it starts, the process could be rapid. The education of the masses in this matter, thanks to cryptocurrencies, is more advanced now than it has ever been. If bitcoin soars to $20,000, $50,000 and more, millennials round the world will understand that the dollar, or their local currency, is going down. The rush out of fiat bank deposits into crypto on its own could easily precipitate a widespread currency and systemic crisis.
This is not to advocate buying bitcoin, or gold for that matter. It is just to warn of the approaching end of the road for unbacked fiat currencies at a time when governments themselves face bankruptcy.
There is a well-known saying, that governments can’t go bust. Don’t you believe it: it depends on fools continuing to place value in their fiat currencies. We can begin to see that end in sight.

OpenAI o3 Might Just Break the Internet (Video - 8mn)

  A catchy tittle but in fact just a translation of the previous video without the jargon. In other words: AGI is here!