So are you still undecided? Do you want, or do you have, devices in your house that take advantage of machine learning? Get Surfshark VPN at surfshark.deals/undecided and enter promo code UNDECIDED for 83% off and 3 extra months for free! If you liked this, check out Solar Panels Plus Farming? Agrivoltaics Explained: kzread.info/dash/bejne/npuOpM59Xbepdqg.html

@davefroman4700

2 жыл бұрын

There is no such thing as too much storage Matt. The more storage we have? The more nearly free energy we have. There are many combinations of generation and storage that can achieve 100% renewable energy. If you plot it on a graph it presents a U curve. However, For every 10% extra you spend on infrastructure above where you chose to set that line? You will receive an exponential amount of free energy in return. We need to get past this scarcity mindset of the past. Technology has long surpassed it. kzread.info/dash/bejne/goFmtNqQpKmadNY.html

@thegreataynrand7210

2 жыл бұрын

Yes I know wind and solar suck right?

@julianshepherd2038

2 жыл бұрын

@@thegreataynrand7210 well don't buy shares in them.

@shriduttshah2347

2 жыл бұрын

by this point you should quit ui/ux and probably go back to engineering school!

@christopheraaron2412

Жыл бұрын

I find all your videos to be very insightful and informative. Just out of curiosity how much do you know about the application of machine learning to actually help accelerate the improvement of computing technology perhaps some of it on the hardware level but also on the software level so then the computing technology can make better use of its hardware by better software seems like that's also a big part of this too I would think thank you very much!

Semi-related, I vaguely remember a story where a group of pigeons were trained to identify malignant cancers in medical scans. The process was similar to machine learning; they started out with a sample of known pictures, any time the pigeon selected a picture with malignant cancer, it was rewarded with food, thus training the birds to select pictures with malignant cancer. I want to say that any given pigeon had less than 80% accuracy, but when they ran it as a group (presumably something like requiring at least 60% to say yes) the accuracy climbed to over 98%.

@SaveMoneySavethePlanet

2 жыл бұрын

Hmmmmm, all those small chirps that MRI machines make suddenly have a different meaning…

@47f0

2 жыл бұрын

Pigeons have also been tested as the guidance units in early smart bombs, pecking at a video screen to keep a target centered, and they have also been tested by a pharmaceutical company to pick damaged pills and capsules out of a production line. Apparently, they did quite well at those tasks, but for various reasons were never used - as far as we know...

@sophiophile

2 жыл бұрын

@@47f0 Skinner, who developed our early theories of conditioning, worked with the military for bomb guidance. But then computers took off, and obviously were more accurate, scalable, and cost-effective.

@namAehT

2 жыл бұрын

I mean that's basically how humans are trained, positive reinforcement. Though if I were given a cupcake for every task I complete I'd probably see my productivity go exponential.

@JorgetePanete

2 жыл бұрын

First time I learn about pigeons doing these things, incredible stuff

It's really nice to hear something hopeful in these turbulent times. Much love Matt!

@UndecidedMF

2 жыл бұрын

Glad you enjoyed it!

@johnnymonsters9717

2 жыл бұрын

I see what I did there

@visualcontrast

2 жыл бұрын

Nice use of the word turbulent... :)

The problem with using a neural network to determine grid functionality is that they're effectively a black box. If something goes wrong, there's no debugging because we genuinely don't understand how the network works. If you want to make a change, you have to retrain the model.

@animefreak5757

2 жыл бұрын

If the neural network is simply coming up with a better layout, this wouldn't be a issue. If we give a AI control of the grid itself..well...i hope we aren't that stupid.

@Kamikater2

2 жыл бұрын

​@@rogerstarkey5390 the problem with neural networks is that they can't explain you how they come to a conclusion. If a human makes a decision and you find it strange you can ask him and he will tell you "I saw parameter a,d,f increasing so I thought solution i would be suitable to counteract that withour harming parameter g and h." a neural network will just do solution i and you have no idea why (and can't look it up as well). In a complex system this can become a big threat especially if the neural network comes to a conclusion like "the power fluctuation is no longer a problem if I shut down the whole grid" and shuts down the whole grid instead of really tackling the power fluctuation (basically "if I'm not playing I can't lose!") I'm not saying that we shouldn't use them, but that we should be aware of the difficulties that come with using them.

Nvidia: "We build hardware that can help model and predict global warming." Also Nvidia: "Look at this gargantuan GPU we made that sucks up nearly 1,000 Watts!"

@mortenchristensrn2487

2 жыл бұрын

Yeah those 3090's are awesome...

@peterclark6290

2 жыл бұрын

They were claiming only a few years ago they had designed and built CPUs that ran on 12W (the same as the human brain). Challenge them and you got instant free (and unwanted) admittance to the Flat Earth club. TBF: when they design a power supply system and circuitry that does away with resistors it may be possible.

@dragaashenoak5338

2 жыл бұрын

The duality of man

@jordanflinn318

2 жыл бұрын

well they need more energy for their GPUs, so they're improving energy to improve their GPUs. Ez clap

@Robert-qq9em

2 жыл бұрын

@@peterclark6290 A lot of circuitry doesn't use resistors much because they aren't great. Sure they end up parallel to switches, but that's got to do with draining your switches so they turn off, they aren't really lossy. You can get PSUs over 90% efficiency without shopping too hard, which means you are losing 1 watt out of every 10 at full load, and ratings (so minimums) are at like 94% at half load. So if you are busting out a 1kW PSU at half load you are losing 25 watts of your 500W, or getting 475W of power through your system on the low end. That's really not terrible. In fact that's pretty impressive to me because it's transmission line level losses. What's even more fun if that most PSU's are shown to run EVEN MORE efficient on DC backbones, so that 25W turns into 20W or so if you replace AC circuitry with DC. Now all this being said? I know not all PSU's run that lean, but that's just an argument for regulation to me, since regulation has been shown to have the economic benefits in these types of cases. Don't let companies sell 60% efficient PSUs, push the floor up to 80%+ type deal.

One potential concern of machine learning when it comes to other things like designing products that work best for humans is that they can potentially further alienate people who don’t fit the majority model. For example, people with autism struggle in shopping centres because those buildings are designed to encourage echos and pipe music to give the feel of a bustling busy environment. That can cause painful sensory overload for some on the spectrum. Similarly, supermarkets use eye-level shelves to market some of their biggest deals, but those in wheel chairs don’t share the same eye level as those who are able to walk.

@matthewmullin8168

2 жыл бұрын

right, but these factors shouldn't influence energy production data.

@AsciiSmoke

2 жыл бұрын

@@matthewmullin8168 - Yeah. Totally agree

GE's Podcast Theatre did a story called LifeAfter regarding digital twins. Really cool seeing the ideas presented there being discussed as real world tech. I think GE needs to do more of those, as it's a fantastic way to get concepts out to the public.

Thank you Matt! As a data scientist aspiring to work in the field of renewable energy, this video is a gold mine. Do you have any resources of analytics and data science projects that I can tap into and learn from? Not sure if you'll get a chance to see this but, thanks a ton. You're a rock star! \m/

As always, very interesting video. I love your videos, Matt, even when a good deal of the content escapes my tiny mental capability. However, I think the bottom line on energy falls with how we live. We are in the process of adding another 4 solar panels with larger wattage to our small home system. Having the solar has made us much more aware of how much energy we actually use, and that is essentially the bottom line. We have to cut our energy usage by beefing up our living infrastructure, not bugger up the entire beauty of our planet with windmills and endless fields of solar panels. Keep up your good work, Matt. Always enjoyable to hear your take on things.

Machine learning would be a huge help in winter road maintenance, predicting how much snow and ice response will be needed over a large area on the roads how much sand and chemical yeah I sure would be needed for each winter. Currently we only have past weather data to go by and passed expenditures. A lot of money and grief could be saved by knowing how much to allocate for a system and when.

Matt, I was underwhelmed with "Sense" AI service. I had it installed for about 6 months, but it never seemed to identify many of the devices installed in my house. There was just a pretty always on or showed that we had consumed X number of watt hours. This included t6hings like refrigerator and standalone freezer. When it did register a device, I do not think it ever identified it as the actual device. I have since installed an Enphase system with batteries and gave up on trying to identify items consuming electricity.

I am still troubled by the fact that all these high powered renewable companies fail to recognize the fact that wind and solar are two of the least efficient forms of renewable on the planet and can never produce the level of efficiency or continuous base load of HYDRO POWER! Through out the world there are areas of land that are subject to consistently high levels of rainfall on terrain that is ideal for hydro electric dams, ( primarily in sparsely populated deep valleys with over 3 meters of rain per year!) Then there are run of river units that can range in size from small individual units that power only a single off grid dwelling to ones that power a city! All requiring minimal maintenance over the extremely long life span of them compared to that of wind and solar! Then there are wave and currant power generation, along with tidal on numerous coastal sites around the world that should be developed

@christianhorner001

2 жыл бұрын

Hydro does not fit the narrative. Engineers & physicists are the enemy who spew hateful facts instead of lovely feelings.

@mv80401

2 жыл бұрын

Efficiency is less important if you're not wasting 80% of a fossil fuel to get 20% of final energy for actual work. With wind and solar the 'wasted fuel' is free, and the modules start paying for themselves quite soon.

@TheLosamatic

2 жыл бұрын

@john monc… it’s really not that hard to understand, all the solutions had to be implemented around politicians who were politicians only to service the mass murdering fossil fuels industry! Plus the petroleum industry long ago infiltrated the environmental movement first to discredit it then to steer it in a beneficial direction for them to remain viable long after anyone with three interconnected brain cells would see that even going back to dark ages living would be better than what is coming much faster than any country that knows will tell its people! Until the transparency of the block chain shows that having politicians from the bottom third level of the intelligence of the people of that country has made paranoia a top issue hence secrecy is shown to be absolutely necessary! When in fact secrecy is only beneficial to the politicians who’s actions were criminal! Long ago governments needed to move away from politicians and much more to managers drafted for say just five year terms. Managers picked not like the orange buffoon did stating that he would drain the swamp when in fact he so hated that a black man was vastly smarter than him, vastly more capable, his narcissism made him undo any/everything Obama. Even when he understood 0% of something then appoint a complete moron who knew 0% of what he was being appointed too the number of voters for such dangerously lacking intelligent appointing, reached record numbers! Point in question was to the department of energy which neither moron understood oversees every aspect of nuclear in the country. Obama’s appointment, one that only a fool who understands nothing but racist hatred of one he knew he could never be near his level of intelligence so tried to erase anything he did, one of our top nuclear physicists. Again the most ignorant among US when asked blindly assuming no skin color would immediately question dumpster’s ulterior motive for most all the things he did, and when confronted with color even the most racist will know how utterly stupid it is to blindly run the whole country like that when it’s becoming so apparent how important it is to appoint from the best and the brightest we have! Until all of the Earths people understand, and stand for transparency, we will continue to see that a country that maintains secrets cannot possibly govern by rule of law, fucked mankind will remain!

@MLRTrytonix13

Ай бұрын

Environnemental consequences is the reason. Living in Quebec, every home operate on electricity that has been produced from dams. The thing is, more dams mean more chances of environnemental catastrophes (leakage or failure) which are devastating on the wildlife. Wind and solar energy don't do that.

Just like to comment that I really appreciate your approach to sharing information. Although there is always a certain unavoidable bias, you stride to give both sides of every argument and let the audience make up their own opinion about each topic you cover. So, thank you and keep up the amazing work!

My first encounter with machine learning was in the mid 90's while installing a 230KV circuit breaker for a capacitor bank. The difficult part for a cap breaker is dropping the cap rather than energizing it. The Breaker had 3 independent poles. The goal is to open the contact at the zero crossing for each phase. So 3 phases opening 120deg apart. The controller would sense how close to zero crossing A phase was on first operation and slowly adjust opening time in consequent operations. So the machine learned how long it took the breaker to open once the open signal was received. Then adjust it to near zero without further input from operators.

Videos from the clean energy creators like this are such a nice thing to look forward to. Thanks

@UndecidedMF

2 жыл бұрын

Glad you think so!

If people invested in passive homes and then did this as well and we did this for an entire nation then I truly wonder how much power would we actually use in America. Biggest issue will be the factories and corporations since they are the biggest offenders anyway.

@MichelDerome

2 жыл бұрын

Even for them it is profitable to make factory more efficient in energy consumption. The more it will cost to use energy one way or the others they will seek way to reducing costs by adopting more efficient building and so on. My point is that it is not to not consume energy but to makes more with the same amounts of energy.

@1voluntaryist

2 жыл бұрын

@@MichelDerome The two concepts, 1. Nega-watts, and 2. Efficiency, are not at odds, not mutually exclusive. Given that, what is your point?

@MichelDerome

2 жыл бұрын

@@1voluntaryist Never said it is at odd. My point was even big corporation having interest to becoming more efficient regarding the use of energy and to make the transition to go faster is to make corporation paying the real cost if fossils fuel. Right now the energy cost of producing electricity has already increased because of the use of fossils energy that are directly connected to the international market. Efficiency is everywhere but even special interest are swing the public opinions in their favor. Hydro Québec was supposed to send renewable energy with their power lines were 80% of they required work to clear the forest for the corridor but the fossils energy organised a public consultation and spent dozens of millions of dollars in publicity and desinformation we see the same regarding renewable energy. So everywhere in every country USA, Canada, Europe they are attacking the coal or petrol energy are using divers tactics to slow down the renewable energy or even penalising the users of solar energy even in Australia 🇦🇺 were they are charging fees or reducing the prices for buying back energy from the consumers. Efficiency is required everywhere but cannot be seen as the only solution for the grid. California is now making harder in the sunshine state to use solar energy. In Australia some solar users decided to buy battery as in some cases it make sens to stop feeding the grid and keeping theirs solar electricity. I would not be surprised that someone will even make it almost impossible to buy solar panels at one point or limit so much the numbers of solar panels that it won't be profitable to use them.

Medical researchers where using machine learning to check if moles where cancer or not. They feed alot of pictures of moles though the machine and it ended up with a 100% accuracy. As in it picked up on every cancer mole. They looked into it and it ended up that the machine was not even looking at the moles, it was looking for a tape/ruler as all the cancer moles had one in the picture.

Excellent Analysis, Deployed Worldwide Through My Deep Learning AI Research Library… Thanks Matt

@robertfoertsch

2 жыл бұрын

kzread.info/head/PLG7EoBMUD1JwbD5-MQpFRGvadGWtf-4yD

@UndecidedMF

2 жыл бұрын

Thanks for watching!

Neuroscientists eloquently summarize Hebbian learning as: Those that fire together, wire together. Those that fire out of sync, lose their link

Great thought provoking content as always Matt, thanks. When it comes to all the sensors that are being deployed to monitor and measure our energy usage as well as the performance of devices at work as well as home, I often wonder what the total resource impact is of creating those sensors. From mining the raw materials to creating a product and powering it on, I wonder what's the total impact on the planet. Not much for a couple of sensors I'm sure but multiply that many times that and it might not be insignificant. Do the same for an electric car or a solar panel. How long does it have to be running for before the item starts to reduce its impact on the environment? Just something I've wondered

The solar lab at my school has transitioned to almost all data science specialists to try and advance PVs faster

just talk to a business chatbot to learn how far ai has progressed.

An excellent approach to renewable energy, covering both the opportunities in production as well as consumption and even the maintenance of our equipment. Eye opening!

Really enjoying all your shows there extremely detailed and challenging! Thank you Like to see shows on solar photovoltaic design comparison like 12/24/48 volt system and show how solar in series is more affected by shading and how fire chances increases with increasing voltages. I’m clearly in Edison camp with a touch of Tesla but not to forget Franklins lessons as well . Again thanks !

Fascinating application area for machine learning for some areas but not a fan of others. I spoke with a utility billing software engineer 15 years ago who told me 30% of my power bill was the cost of the billing/metering software to calculate my power bill - grrrr. Smart thermostats are great & all but I like the idea of ML scanning different compounds for 'normal' temperature superconductivity - that would be a real win.

A few years back, I pseudo-coded how I could set up a machine learning driven system for when to turn on and off a fan in my house at night based on predicted weather. I never got around to it, but this just reminded me that I should have pushed more on that.

@Deathington.

2 жыл бұрын

You don't need machine learning for that.

@kerradeph

2 жыл бұрын

@@Deathington. No, I don't need it, but if I wanted to optimize when to turn off the fan then I could get a little bit more out of it.

@NoNameAtAll2

2 жыл бұрын

@@kerradeph that seems like it's solved a lot better with temperature sensor...

@kerradeph

2 жыл бұрын

@@NoNameAtAll2 Yeah, I was planning on having internal sensors that I could tie into a Raspberry Pi monitoring system along with an outdoor weather station. I was expanding on the possible uses of those since I would be collecting and storing that data locally anyways.

I keep watching your videos about renewable energy because that's what i´m studing about. I wish my friends could to understand english to follow you content.

Hey Matt, there seems to be a categorical blame of landfills whenever there is some sort of discussion of our environment. However, there is never any proof, data, or arguments presented to back up the assertions. I believe that the evils of landfills are grossly blown out of proportion. For example, what's the issue of burying discarded plastic in a landfill until it breaks down? Plastic came from the ground and now it's going back. Are we actually "clogging up and filling up" all of our landfills, or is this a fallacy? Can you please do a video on the subject? Check out the episode of Penn & Teller Bullsh*t to see what I'm talking about.

Having installed a Sense monitor in my home about 2 years ago, I've been disappointed in its ability to identify individual devices...an ability that Sense markets pretty strongly. The system still classifies about 60% of my consumption in the catch-all categories of Always On and Other. Not the fine-grained detection I'd hoped for.

@animefreak5757

2 жыл бұрын

the finest it can get would be per circuit, with maybe a little AI sort of thrown in to detect certain loads based on power factor or something. If there a labelling process for this monitor? or does it just guess at what each circuit is for?

@deanwight

2 жыл бұрын

@@animefreak5757 Sense has current transformers only on the main cables, not on individual circuits. There is no ability for users to "teach" it by, for example, saying "I just turned on the dehumidifier". It notifies the user when it has found something it believes is unique, and gives it a name like "Motor 3". You can tell it to notify you when Motor 3 turns on or off. So hopefully you can determine what it is, and name it accurately in Sense. But many routine loads, like my outdoor lighting that's on a timer and runs every day, have never been identified by Sense. I don't know what combination of voltage, current, power factor and etc. are used by the system.

@animefreak5757

2 жыл бұрын

@@deanwight well, that's unfortunate. now that i think about it, per circuit would actually be a bit complicated due to differing breaker layouts and such. I've seen devices that are separate and you can put current clamps on any line you want. In theory you could even run signal wire out to individual devices.

@pazi402

2 жыл бұрын

Have you looked into the smart plugs made be Emporia? You can configure the plug to associate with the circuit the device is connected to and the software will balance out the usage separate from the circuit. Sell the Sense controller and switching to Emporia might be worth it, you can export the data you currently have to and Excel database to compare.

I like the machine learning that used cameras watching mirrors to make the solar mirror array stay at highest efficiency throughout the day, so simple and smart.

I think machine learning in renewables is definitely coming and it’s happening faster than people might think. Enphase has production and consumption CTs in their platforms and Span has CTs on each individual circuit. Making homes smarter on energy use is a very, very good idea. I work in the renewables sector and smart technology is coming very quickly to the industry. The only downside is we now need to protect low voltage CT wires otherwise it can lead to some faulty readings.

Matt IS the best !)

I find it really interesting how this music makes me feel. The music is the same every time and feels like it fits in with the "Corporate" genre of soulless music, but every time I hear it I get excited because of the content of the video.

Thus far in 2022, the Texas ERCOT grid wind has provided 31% of the electricity there (and another 5% from solar)... largest electricity producing state in the US by a good margin over FL... add in some 1 and 2 or 3 % improvements along the way, add battery storage and demand side signalling to back off briefly ... it's going to make a big difference for the portions of the year when wind is so active there. I see these fields of solar panels at industrial production level all nice neat rows, and i know why they do that ... maybe a more scattered orientation approach matched to TOD generation requirements is more in order and some of the AI / ML modeling and analysis can take geography and do the best results layouts... Interesting video.

You and Just Have a Think are awesome! How about a podcast discussion with you and Just Have a Think, for 3 hours? I would listen to it while building speakers.

We have had solar panels for 20 years. We got a $26 dollar rebate for the year last year (and our electric company has the highest rates in the nation). My question is, we have a Chevy Volt. Should we charge it during the day or in the middle of the night? We have been charging at night for the lower rate but if we are making electricity in the middle of the day, would it divert to the Volt if we charged it then, or will it first go back to the grid?

Its a start but I will be more impressed when the AI advises to move the solar power to LaGrange point 1 and tells us how to get the power back to earth

@JasonPurkiss

2 жыл бұрын

@@rogerstarkey5390 thanks, but that why I want AI to solve it :)

I'm scheduled to get a Span panel but I need to figure out how to accommodate 43 breakers with a 32 system. What do I want to not battery back-up or monitor individually?

I've read that with machine learning you have to be very careful about the datasets and goals. Some were inadvertently biased by showing images of people of only one race. Another was given the goal of learning to 'walk' and 'learned' to flip on its back to meet the goals set by the programmers. Often the programmers are surprised by the unexpected results.

Nice for those with the resources to implement these cutting edge technologies. Now we need to mainstream them. I just had a dumb whole house surge protector module installed my breaker panel and that is cutting edge for me. A basic consumer. Thanks Matt for the enlightening vid!

@iglapsu88

2 жыл бұрын

@@rogerstarkey5390 thanks

Love your work, Matt! But I have to playback at .75 speed to keep up... Maybe you should cut back on the goies.

This is why many excited about seeing Tesla apply their design skills and data gathering experience to HVAC and home energy capture and storage systems. It hasn't been announced publically, but I wouldn't be surprised to hear that Tesla have been working on the technology.

Using AI to streamline efficiency etc. is doing calculations way quicker than humans. Sounds crazy but it's really no different to using a ruler instead of just guessing. A ruler is way more accurate and consistent than a human.

Impressive to say the least. getting a good mesh/distributed grid optimized for efficiency will be such a relief. I noticed something for the first time. the next big industry is TIME. machine learning, and predicting what's next in any system. from chemistry to weather to probably even human behavior to a disturbing degree

In southern California the temps never get really low or high. We have lots of swimming pools. Do you know if pool water can be used for the back side of a heat pump? In the winter the pool water has an average temp of 55*F. In the summer, the temp increases to 80*F.

Your transition into SurfShark does remind me of the dark side of machine learning: It can be used for bad as well as good, which becomes more worrisome as the technology becomes more capable.

does this mean that with predicting weather and adjusting yaw of wind turbines we could make up to 50% gains in energy production from existing renewable infrastructure? Is this being factored into modelling projections for renewable energy production?

Happy tuesdays Matt!

@UndecidedMF

2 жыл бұрын

Happy Tuesday!

quite interesting and good chanell you have, worth a sub for sure. topics are excellent, some discrepencie with pictures and text but good

This makes me think of the movie Terminator. Once the machines control the power supply you can't unplug them. Better hope they learn nice things before they control to much.

Hey, a topic I am familiar with. I think machine learning and AI is the most important first thing to improve. With it all other topics can be handled much easier as you have now a virtual companion who can think with you about other problems and come up with solutions a human wouldn't even think of.

I've been looking everywhere, but I can't find anything on glove phones. the first time I saw it was in an anime (orbital children on Netflix) and the real-life one is a mitten that has a speaker and mic in the pinky and thumb. My question is, do you think that using OLED and conductive fabric would it be possible, or would we have to become bio-augmented (brain implants that project an image along your body for personal use) to make it possible?

Absolute FIRE vid Matt ! Keep kicking all the ass presented & REMEMBER USE BOTH FEET!

Fascinating. Thank-you Matt. So, do we have to change the saying "It takes a rocket scientist" to "an AI scientist" to figure out the most complex problems we see?

thanks. Love Watt-Sun and Watson!!!

I have decided that I am Undecided, thanks to Matt.

Machine learning can improve efficiency thus improves effectivity as well.

@UndecidedMF

2 жыл бұрын

Absolutely

Great for developed countries, probably beyond the reach of developing countries. Where populations are sparse, decentralised mini grids (community energy) will predominate - best to keep it simple. Also, surely as energy storage advances & becomes cheaper (hopefully) doesn't this reduce the need for micromanagement of the grid? There are ways to max energy efficiency (green roofs, shading, building envelopes, double & triple glazing, insulation, passive houses etc) & reducing energy demand. In fact, energy efficiency should always be the first step in managing energy needs and is often the most cost effective step in the process. I'm very nervous about being over reliant on gadgetry because if anything goes awry, then the whole process could collapse like a house of cards. I'm particularly anxious about hackers or terrorism...

@georglehner407

2 жыл бұрын

You want both. Energy efficiency as well as AI optimization. As a side note, I personally doubt that usage of machine learning increases the chance of hackability that significantly - machine learning basically just fine tunes parameters, any system that can be hacked by changing parameters can be hacked just as well if it hadn't used machine learning. Also, machine learning will be applied to both the large as well as the small scale. The problem with machine learning is usually the training process of the model; whereas applying a model tends to go quick and doesn't need a lot of computing power. In this sense I believe machine learning will change the game both for the developed world as well as developing countries.

Thanks Matt. 👍🏻

Fantastic & rich tech content

Virgin's optimization finally! Undecided with Matt Farrell and Ticker Symbol YOU! You too absolutely should have a sit-down one on one!

Matt, do you think it is more efficient / better, to have people install renewables (solar panels or smaller wind turbines) on their homes, versus massive solar / wind farms? I'd like to install them on my home someday, but am curious about your thoughts... Hope all is well and have a most blessed day!

Good job. Public could use more information regarding difference between ML and AI.

Man. Clearly. Digital Twins has shifted in scope. Thank gosh. Was stressing me out. Great vid as always.

@UndecidedMF

2 жыл бұрын

Glad you liked it, thanks!

Machine learning so that an appliance is replaced or repaired before it dies sounds very cool particularly for tradies that don’t appreciate call outs after hours when the owner finds appliance has died 👍

Nice bro good info

Maybe add electronic data/information sharing systems to speed up the massive input of data needed for AI to analyze accurately and relatively quickly.

Amazing video Matt, Greetings from Spain

4:45 I think the third one that is usually mentioned is Reinforcement learning.

The main problem I have with machine learning is the people who implement it. Usually the people implementing it fail to set hard limits on it, letting it learn to ignore the user preferences. Or they will let it learn for a while and then reset it over and over. I used to think nobody would be stupid enough to program the gray goo "paperclip machine" so carelessly, but I keep seeing people do that with machine learning, letting it run free and maybe going occasionally to correct the biases when someone complains the machine learning is acting a little too obviously evil.

I am actually using the sun forecast in my smart home integration.

10:22 this simulation seems off, wouldnt the further behind the more wind controls the wake? So should look like a curved smoke?

Background music kept making me think an ice cream truck was driving by outside 🤓

Another great video Matt! After spending the entire day crying over my portfolio I wake up to this video! A good start to the day!

Hey Matt! Really >>>LOVE

you talked about 3 types of machine learning. but, what if you have all 3 types working on the same input? how similar are the outputs? what can you learn from the differences in the outputs? can they train eachother, with those outputs? would it be better to have a 3 output system, to have the human review, as needed?

Did you heard of the Idea of the vertical axis Windturbines which a paired together to generate more electricity? It uses the same principle what fish use

Love your channel! I’m also a fan of Matt

Interested to see if this could be used in the Pulp and Paper / lumber industry where huge volumes of wood waste is created.

Thanks for your knowledge sharing

This is exposing a very big weakness in Renewables the cost to produce not only is The Upfront costs exorbitant the production cost now will be high

becoming reliable on digital infrastructure gives more vulnerability to failure and remote sabotage, I'd say keep systems as mechanical as possible even if it means trading in efficiency for long term security

"Mr. Watt-Sun, come here!" -Mr. Bell

Australia (southern leading) the way with WIND & SOLAR - do research on that.

Please talk about Ecosia they are a search engine that plants tress

well, here they promised us to be able to see usage, when they were putting electric smart meters in. several years later, still no end user functionality whatsoever. they got to save loads of money by not having meter reading staff, and we got always increasing prices. also our deregulated market produced lots of bill printing "power suppliers" who dont supply anything but often innacurate bills. we were promised competition, lol.

Every time you said "digital twin," I thought of Microsoft Flight Simulator - because the developers often referred to their model of the Earth as Earth's "digital twin."

Very interesting video. From my perspective within the industry I am a bit puzzled. Not seen anything like that yet. Even with Siemens GAMESA. Or the he names you put forward in this video.

@mathieusarran7833

2 жыл бұрын

Not sure AI will do much about wake and yield analysis but I believe that could be a positive feature for grid operation… although this is not too complicated. Instead of artificial intelligence, we just need common sense and consume less energy.

@mathieusarran7833

2 жыл бұрын

Do not wait for the technology développés by others help you…. Do your share first

One thing that you haven't mentioned that is one of the big problems with machine learning is database bias. Granted in engineering it is easier to noticed that you might be missing data. But it is one of the big problems that well since the most data we have is focused on the "developed" world the learning algorithm will be biased towards the area of that too. As soon as one is aware of that problem one can attempt to fix it of course but if it is not mentioned it might be overlooked as a potential source of problems.

Wish you explored whether self driving cars also reduce overall emissions. I know that’s not really an objective for those models but someone is thinking about it right?

Great intro on machine learning man. Thanks 👍

@UndecidedMF

2 жыл бұрын

No problem 👍

The one area I would like to use is the electricity data on large appliances. I'd like to see when the machine is starting to break down before it actually fails.

@animefreak5757

2 жыл бұрын

seems to me like the most common failure on modern appliances isn't the main load (such as heater element or motor) but rather the circuit boards that are in them. I'm pretty skeptical that such a thing would be detectable just by the amount of power going into the device.

Gracias por tanto. Un saludo.

Please consider doing an episode on aerosol shock and global warming. As we remove aerosols from the atmosphere potentially accelerating warming as the energy they reflect now passes to earth.

also don't forget that there's also GOFAI (good old fashion artificial intelligence) that runs on normal math and decision trees made by programmers (not machine learning), and it is also being incredibly useful in almost all situations where a computer is needed.

Matt, I remember a few years back Intel used to sell a USB neural networks chip for home projects. Can you provide any new home or business kits and tools for AI home lr start-up projects, as a follow up on this video?

@UndecidedMF

2 жыл бұрын

Interesting. I'll have to dig into that.

@MichaelDorchain

2 жыл бұрын

You can check Coral USB Accelerator, it might fit your needs

Matt, do you have links for downloadable clean energy software?

Is this M.L. being used at Tesla micro grids, e.g., connected PV roofs/batteries?

@9:51 Perhaps... the more efficient a turbine layout is, the more likely birds will get hit... a bird flying a stright path might not be able to avoid blades if a region is blanketed by offset turbines.

@DubDTube

2 жыл бұрын

@@rogerstarkey5390 I stand by my "perhaps", in that what I suggest is likely true. Clearly sounds like we need to do something about cats. They scratch, and infect the brain with mind controlling microbes. Google that.

Mind-blowing!