Mediphany

mediphany

Cannabis is gradually coming out of the closet. In Germany, the UK, France, Italy, Argentina, Australia, and some other countries in Europe and the Americas can already use it for medical purposes. In the Netherlands, Georgia, Spain, Mexico, Brazil, South Africa, and other countries, there is no criminal punishment for the recreational use of marijuana in non-public places. And in Canada and Uruguay, the use of marijuana in all forms is completely legal. In the US, the use of cannabis is prohibited at the federal level. However, 35 states and the District of Columbia (DC) its use for medicinal purposes and 15 states – for recreational use. have legalized There is a growing number of US states that continue to move towards legalizing cannabis. States considering this move include Missouri, North Dakota, Idaho, and Wyoming. Estimating the Size of the Legal Cannabis Market In 2019, the volume of the legal cannabis market was estimated at $28 billion, according to . Even if everything remains as it is, by 2028, the market capacity $40 billion. In case of medical marijuana legalization in Europe and at the federal level in the United States, it could grow to as much as $55 billion. Barclays Bank will reach According to Grand View Research , the international market for pot and CannaTech products will reach $66 billion by 2025. forecasts According to recent forecasts by , by 2024, the marijuana market will grow to $63.5 billion, demonstrating a 21% growth YoY. Naturally, investors cannot avoid this trend. Statista The latest cannabis legalization trends spark an interest among entrepreneurs looking to tap into this new promising market. A new wave of technology startups hit the market with their digital solutions that aim to make the process of cannabis purchasing, delivery, and consumption as seamless and fast as possible. Each of those startups dreams of becoming the next "Uber" for marijuana in their niche. What is Cannabis 2.0 and How It’s Driving Market Growth Before cutting-edge software and hardware technologies got involved in cannabis growing and distribution, we referred to Cannabis 1.0 (e.g., early stages of legalization and strains’ distribution through dispensaries and coffee shops). Today, when technology takes center stage, we’re speaking about the Cannabis 2.0 industry. Cannabis 2.0 solutions cover a broad range of pot products, derivatives, and services, including but not limited to: Cannabis consumption devices such as vaporizers, vape pens, electronic joints, dab rigs, oil rigs, and others; Cannabis-infused food products for oral consumption such as edibles, tinctures, sprays, etc.; Cannabis dosing devices such as pens, vapor meters, and so on; eCommerce platforms and delivery apps; Systems for indoor and outdoor weed growing; Equipment for the industrial cultivation of cannabis; Breathalyzers – devices that determine THC's level and other active substances in the patient's body by analyzing their breathing. Now let's dive deep into some of the most thought-provoking CannaTech use cases. CannaTech Use Cases Across Verticals and Niches Cannabis 2.0 market consists mostly of (THC), (CBD), and (CBN) infused products, their derivatives, accessories, and supporting products. tetrahydrocannabinol cannabidiol cannabinol In addition to THC (a psychoactive substance that creates the "high" effect), hemp plants contain over 100 other active ingredients. More than 25,000 different products can be made with hemp, including food, textiles, construction materials, ethanol, and biofuels. No surprise that the diversity of cannabis properties and the growing infrastructure around it have been grabbing entrepreneurs' and investors' attention worldwide for several years already. So how are different industries and domains taking advantage of cannabis legalization and mass adoption? 1. Medicine and pharmaceuticals Weed has been widely recognized for its psychoactive and relaxing effects for decades. Numerous studies have confirmed that CBD helps with epilepsy, neuropsychiatric and anxiety disorders, cancer, etc. Recently, the US Food and Drug Administration (FDA) has to be used officially to treat the two types of epilepsy. Cannabis is also used as a substitute for opiates for chronic pain. One marijuana helped 80% of those surveyed relieve chronic pain, and 50% of them had used opiates before. approved CBD study says According to , over 3 million patients in the United States regularly use prescription marijuana. Marijuana Policy Project According to a study by the University of Georgia, the pharmaceutical industry $4 billion in cannabis sales annually. Many pharma companies are considering incorporating marijuana into their selling strategy. generates is a world leader in the development of drugs containing CBD. The company was the only one to have received FDA approval to use their drug Epidiolex for the treatment of epilepsy. GW Pharmaceutical Canada's recently announced a partnership with Sandoz Canada (part of Novartis) – the first contract between a large pharmaceutical company and a marijuana manufacturer. Tilray 2. Cosmetics and products for a healthy lifestyle CBD oil is a promising new product in the industry. It doesn't make you high, but it can relieve pain, anxiety, and depression. The oil also provides an anti-inflammatory effect and helps with acne. In addition, athletes can use it to ease inflammation after exercise. It also helps in building muscle mass. The popular US brand makes candy and body lotions from cannabidiol to help treat muscle, joint, and skin pain. Potentially, oil-based lip balms and mascaras can be created to foster the use of natural cosmetics among users. Lord Jones 3. Manufacturing and packaging The fibrous structure of the Cannabis Sativa plant stem makes it one of the top contenders for the market dominance in renewable materials. Hemp packaging materials are a . It can also be used to produce paper, clothing, and even building materials. Hemp seed oil is used in varnishes and paints. safe alternative to plastic The packaging of CBD-based products is strictly regulated by law. Among the : requirements are a special cover that prevents access to children; indicators of unauthorized access, non-transparency, etc. Many new developments have emerged to meet specific market requirements. A lot of people use bags specially designed for storing cannabis. And also ziplock bags, cans, slide boxes, tubes of different colors, and small boxes, like those for contact lenses. That's a huge separate market that will grow bigger as cannabis legalization continues gaining steam all over the world. The benefits of industrial use of hemp do not end here. Cannabis is excellent at retaining carbon dioxide and detoxifying the soil, helping combat global warming, and so on. Therefore, the ecological potential of cannabis makes this industry extremely attractive for both entrepreneurs and investors. Plastic is inexpensive and versatile. But the fact that it is not biodegradable is a significant threat to the environment. The world's leading economies allocate $80-120 billion each year to solve this issue. Hemp can become a healthy alternative to plastic and can be used to make everything from bottles to household appliances and toys. , an Alaska-based developer of sustainable packaging, has created a hemp-based bioplastic. Together with their permanent partner , they can produce more than one billion units of various packages per year. Best Practices Packaging Penta5 is also looking to replace plastic with more sustainable material. They have already invested $150 million to research the use of hemp-made plastic in its products. Lego Hemp is a genus of plants in the cannabis family with low levels of THC. The potential of hemp can be harnessed to make paper. Up to 400 million tons of paper are used globally every year; on average, 17 trees are needed to make one ton. One hectare of cannabis can produce four times more paper than a hectare of trees. In the construction industry, has already become actively used. It's a light, cementitious material made from the hardened parts of the hemp stem, lime, and water. It can be used as concrete in modular building blocks. Hempcrete can replace oriented strand boards (OSB), drywall, and thermal insulation. In the US, the first 50 hemp homes were built in 2010. In Washington, Hempitecture remodels homes, while Left Hand Hemp in Denver builds campfire homes using hempcrete. hempcrete A Ukrainian company installs heat insulation in houses using hempcrete instead of traditional eco-unfriendly and toxic materials. Hempire 4. Agriculture and food To produce a ton of finished textiles from hemp, you need half the land than if you make them from cotton. Hemp can grow in different soils and at different temperatures. The plant is drought-proof, so expanding on an industrial scale will help conserve water resources. Hemp plant products containing CBD have anti-inflammatory and cleansing properties and are used extensively in the United States as a dietary supplement to improve sleep and appetite. According to a new by New Frontier Data, in 2020, US sales of cannabis-infused food products have approached $3 billion. report The (official cannabis distributor in the province) is offering a total of 59 new CBD-based products in Canada. A wide range of new products should attract consumers, including those who do not smoke it. Food products will also have much higher profit margins and help Canadian companies cross the profit margin. Ontario Cannabis Store A group of brands has invested in the renowned startup , whose shares rose from 10% to 38%. Beer brands are also interested in developing special cannabis-based product lines. In California, announced two types of psychotropic soda. Canopy Growth Lagunitas Brewing Company Tobacco and alcoholic beverage producers are trying to create new marijuana products to stop their markets decline. Another cohort of cannabis tech startups is specialized in solutions for indoor and outdoor weed growing. The cannabis legalization movement has made it possible to grow marijuana at home in many countries around the world. Canberra, the capital of Australia, has recently voted to legalize the personal use of cannabis for recreational purposes. It implies that residents have the right to grow two bushes of weed at home. Thailand is also currently considering a bill that would allow each household to plant up to six bushes of cannabis. This trend is driving the demand for home-grow systems whose market has attracted both traditional equipment manufacturers and innovative startups. For example, a couple of years ago, Cloudponics introduced a specialized module for growing cannabis. Leaf raised $4.5 million in May this year to develop a plug-n-plant system that monitors the growth of cannabis bushes from a smartphone. Hydroponics is a new trend that many CannaTech companies are looking to tap into. It's a horticulture and hydroculture , without soil, by using mineral nutrient solutions in an aqueous solvent. method of plants growing Hydroponic systems can accelerate marijuana growth significantly, which helps reduce time to market and increase the yields of strains. Brands specialized in the design and production of hydroponic systems include , , , and many more. HTG Supply General Hydroponics PowerGrow Systems 5. Fashion Hemp fibers are used to make clothing, durable textiles for commercial and industrial purposes, and strong ropes. Unlike cotton, hemp is non-soaking and has antibacterial properties. Today's consumers are willing to pay up to 20% more for eco-friendly products, which is why major suppliers and niche brands are heavily betting on hemp clothing. Hemp is appreciated for its durability and economy, but it doesn't dye as well as cotton. Hemp fabric is similar to linen, but it itches a little when touching the human skin. Both mainstream and indie fashion brands such as , , and produce clothing made from hemp fabric, which has a texture similar to jeans or wool. EnviroTextiles Patagonia Bad Decision Adventure Club 7. Biofuels and energy Hemp produces oil per hectare than soy – the only plant grown commercially for biofuel in the United States. four times more Researchers from the University of Connecticut have discovered that hemp can also be used to make biofuels. Scientists plan to build grant from the US Department of Energy. It will be able to produce up to 757 thousand liters of biofuels per year. the first plant using a $1.8 million 8. eCommerce and delivery In line with the rapidly growing cannabis legalization trends, online retail portals pop up to enable the sales, distribution, and delivery of various hemp-based products. London-based is one of the leaders in this space. The company is building "Europe's leading cannabis distribution platform," according to its website. It's more than just an eCommerce platform; in fact, it's an entire ecosystem of cannabis-infused products and CannaTech solutions. Besides a curated online marketplace, Alphagreen has a life science division called . It assists European cannabis suppliers with EU GMP certification and import, batch release and product registration, warehousing and distribution, support in marketing and branding, and more. Alphagreen Alphagreen Med In the United States, is the leading cannabis delivery app created back in 2014. Backed by some $166 million in funding, Eaze is the first CannaTech startup to have raised big money in Silicon Valley. Eaze According to , Eaze is trying "to pull off a "verticalization" pivot, moving beyond online storefront and delivery of third-party products into sourcing, branding and dispensing the product directly." TechCrunch One of Eaze's most promising competitors is , a CannaTech startup that delivers marijuana in American 18 states. The company raised $15 million from Gron Ventures and several other VC funds in September this year. Among Dutchie's investors is Jon Oringer, founder and CEO of Shutterstock. Dutchie Venture Investments in Cannabis 2.0 Although the total number of cannabis venture deals in 2019 decreased compared to 2018, the volume of investments has almost doubled. PitchBook shows that in 2019, there were 307 VC deals with 277 CannaTech companies, totaling approximately $2.62 billion. data The leading VC funds in terms of capital investment in the cannabis industry in 2019 and 2020: Phyto Partners (Boca Raton, FL) – 23 deals; Poseidon Asset Management (San Francisco, CA) – 21 deals; Entourage Effect Capital (Dallas, TX) – 20 deals; Arcadian Fund (Beverly Hills, CA) – 20 deals; Altitude Investment Management (New York, NY) – 19 deals; Salveo Capital (Northbrook, IL) – 18 deals; Casa Verde Capital (West Hollywood, CA) – 13 deals; Merida Capital Partners (New York, NY) – 12 deals; Big Rock (San Francisco, CA) – 12 deals; Cannabis Growth Opportunity (Toronto, ON) – 12 deals. Key Risks Associated with CannaTech Investments There are clear risks of investing in CannaTech today: - Political and legal risks (the upcoming change of the US president and policies in this segment, restrictions on funding cannabis producers, etc.); - Overestimation of supply over demand (those who use marijuana illegally will continue to do so, and there may be too much supply on the market due to drastic legalization). It’s also worth considering that, according to analysts' forecasts, many cannabis producers will simply not have access to reputable exchanges (e.g., NASDAQ), which provide access to really large capital. This fear rises because most of the producers do not have sufficient capitalization and will not be able to trade on the stock exchange. At the same time, the purchase of such shares in OTC trading promises serious risks. Exchange-traded marijuana producers (most of them are Canadian companies), albeit worth billions of dollars, have lost significantly in capitalization in 2020 compared to 2019. As of October 2020, the capitalization of Canopy Growth Corporation fell to $5.79 billion, Cronos Group - to $1.87 billion, and Tilray - $ 692 million. According to , the market is already oversaturated. The capitalization of public companies grew in 2018 on expectations of the legalization of marijuana in Canada. However, the market's legal demand remains limited, and a new growth factor is needed – the market largely depends on cannabis legalization at the federal level in the USA. Cannabis Legale However, not all companies experienced a decline in 2019. At least two CannaTech players felt like rock stars, namely and . Valens Groworks Medipharm Labs Based in Kelowna, British Columbia, Valens is a vertically integrated company specializing in cannabinoid and terpene extraction systems for CannaTech 2.0 products. Traded in both the United States and Canada, Valens posted more than 112% gains in shares in the US OTC market last year while yielding just over 105% on the TSX VIX in Toronto. Ontario-based MediPharm Labs specializing in refined cannabis concentrates gained just under 128% in the US and more than 120% in Canada. Many venture capitalists have turned their attention to the marijuana market since its legalization in California in 2018. The California market is set to become one of the largest in the United States. In 2019, its volume exceeded $5 billion. Nonetheless, since cannabis has not yet been legalized at the federal level, large American funds are not interested in the market. Marijuana is still a risky market for VCs and investors. Banks refuse to open accounts for companies that work in this area. Coffee shops selling weed are only allowed to use cash, which, again, cannot be deposited into a bank account and poses a lot of inconveniences. Final Thoughts If you're looking to tap into the fantastic new opportunities that will unfold as more countries legalize recreational and/or medicinal use of weed, you should hurry to start building a strong CannaTech brand now. In case of mass legalization, you will be competing with alcohol and tobacco corporations (many of whom are already investing heavily in Cannabis 2.0 R&D) and innovative startups like yourself. Which CannaTech startups do you consider the most promising? What startup would you launch yourself? If you're a startup working on cannabis innovation and you have a cool project/product to pitch – reach out to me on LinkedIn and I'll be happy to feature you in my next article or interview.

Cannabis 2.0: “Honey is Sweet, But the Bee Stings”

What’s the biggest difference between an AI model and a human brain? Over time, myriad answers have been given—the brain is more energy-efficient, more multifaceted in its media of input, and also chemically enabled in addition to being electrical—yet the human brain’s most important feature is its amazing plasticity. If a patient’s body part (like fingers, a hand, or even entire limbs) is severed, the neural sensorimotor region corresponding to that body part, now devoid of a nerve ending to connect to, will almost instantly start adapting, with the neurons “switching” to help other nerve centers in controlling other body parts. Plasticity also helps humans ingrain ideas and skills: as the saying goes, “neurons that fire together wire together”. Muscle memory and near-instant factual recall are two plasticity-enabled parts of our life that we could never live without. For decades, scientists have failed to come up with a similar function in AI models—until now. On June 12th, a team of MIT researchers published a groundbreaking research paper demonstrating how an AI system can in fact utilize human-like learning processes to improve its own performance on benchmarking tasks. In this article, we explore the moral and technological implications of the so-called Self-Adapting Language Model (SEAL), the world’s first self-evolving AI. other improve its own performance Imperfect Learning Of course, AI models using the Transformer architecture were still able to learn certain tasks, yet the few methods available were not quite autonomous and far from efficient. Perhaps the most notable way to train a model to perform a certain skill—like translate English to Chinese or do trigonometry problems accurately—was to use a process called Supervised Fine Tuning, or SFT for short. This method worked a little like this: Identify the exact task you would like to perform SFT on. As an exemplification, let us assume the example of generating modern song lyrics.
Gather high quality examples in the form of (input, output) pairs. For our example, an obvious yet controversial way to do this is to simply use song lyrics scraped from the internet and pair them up with rough summaries of the contents and characteristics of the songs.
Perform SFT on the model. This is usually done through a process called Gradient Descent, the technical aspect of which I cannot adequately explain in this article. Over a large number of training iterations, this process alters the model’s weights such that it is able to produce something similar to an output (the actual song lyrics) given its corresponding input (a specific description of a song). Identify the exact task you would like to perform SFT on. As an exemplification, let us assume the example of generating modern song lyrics. Gather high quality examples in the form of (input, output) pairs. For our example, an obvious yet controversial way to do this is to simply use song lyrics scraped from the internet and pair them up with rough summaries of the contents and characteristics of the songs. Perform SFT on the model. This is usually done through a process called Gradient Descent, the technical aspect of which I cannot adequately explain in this article. Over a large number of training iterations, this process alters the model’s weights such that it is able to produce something similar to an output (the actual song lyrics) given its corresponding input (a specific description of a song). Gradient Descent Gradient Descent For all its intents and purposes, SFT did work, remaining a tool within an AI developer’s repertoire to catch specific safety lapses or improve an AI’s performance on specific tasks. Unfortunately, the very nature of SFT meant that the process was inflexible and expensive, often requiring a moderately large quantity of high quality data specific to the field of responses being tuned (e.g. Mathematical reasoning, Grammatical style). Although many research papers have proven that traditional SFT can be performed just as well using synthetic, AI-generated data, SFT remains a tool to be used with caution, since altering model weights may negatively impact a model’s performance in other types of exercises (a model improperly fine-tuned for mathematics might, therefore, suffer a trade-off for essay writing). Inklings of Evolution Note: The information in this section is largely paraphrased from the June MIT study “Self-Adapting Large Language Models” by Zweiger et al. Note: The information in this section is largely paraphrased from the June MIT study “Self-Adapting Large Language Models” by Zweiger et al. Self-Adapting Large Language Models Self-Adapting Large Language Models One of the downsides of traditional SFT has always been the human effort involved—SFT pipelines often had to be handcrafted by human AI researchers, even though it usually was an effective way of tuning a specific model to perform slightly better on certain types of tasks. Met with recent advances in synthetic data, the researchers rejected the notion of simply using AI-generated SFT data, going further to ask the question of whether humans could be moved out of the SFT loop entirely. Their answer, the Self-Adapting Language Model (SEAL), is, in reality, part of a larger framework consisting of a pre-trained decoder-only transformer model (the study used two open source models, LLaMa-3.2 and Qwen-2.5B, for separate testing cases), “tool execution” software, and the SEAL network itself, with a shared goal to answer several benchmarking questions (the context) as accurately as possible. The SEAL network does not actually predict and generate the answer to the question—instead, it focuses on performing SFT on the decoder-only transformer model with the goal of enhancing the model’s performance when given that question. In order to do this, the SEAL network is given two major tools: decoder-only transformer model decoder-only transformer model context) model’s Synthetic data generation: By calling this tool, another network will pick up the context (essentially the prompt) and generate SFT pairs. For example, if given a passage about the developmental history of the airplane, one tuning pair might be (“What was the first ever commercial jet airliner?”, “the De Havilland Comet”). Although a question-and-answer format was frequently used, this tool could generate other types of content to better suit the needs of specific problems.
Hyperparameter tuning: As previously mentioned, SFT is a process that repeats for multiple iterations; the exact settings of the training steps are therefore customizable in a process called hyperparameter tuning. By calling this tool, SEAL can initiate an SFT with specific settings (like Learning Rate, # of Epochs (iterations), or the batch side of Gradient Descent), potentially changing how well (or poorly) the decoder is tuned. Synthetic data generation: By calling this tool, another network will pick up the context (essentially the prompt) and generate SFT pairs. For example, if given a passage about the developmental history of the airplane, one tuning pair might be (“What was the first ever commercial jet airliner?”, “the De Havilland Comet”). Although a question-and-answer format was frequently used, this tool could generate other types of content to better suit the needs of specific problems. Hyperparameter tuning: As previously mentioned, SFT is a process that repeats for multiple iterations; the exact settings of the training steps are therefore customizable in a process called hyperparameter tuning. By calling this tool, SEAL can initiate an SFT with specific settings (like Learning Rate, # of Epochs (iterations), or the batch side of Gradient Descent), potentially changing how well (or poorly) the decoder is tuned. Hyperparameter tuning Hyperparameter tuning hyperparameter tuning Now that SEAL has two powerful tools to help the AI model learn, it only needs to be trained on how to use them. At the start of its training, SEAL applies the two tools randomly for each benchmarking question the framework encounters. These self-edits (SEs, as the researchers called them) will generate contextual, but not verbatim, fine-tuning data within the topic of the prompt and alter the original decoder-only model using the aforementioned hyperparameter tuning steps, making the network produce a different output than before. However, there’s a catch. The researchers didn’t simply change the original model (denoted as θ) directly using SEAL; instead, they made a copy of the proposed changes and incorporated them into a prototypical transformer model (θ’) separate from θ. The training process now goes into an “inner loop”, consisting of the new model θ’ as well as the original benchmarking question. If the model, in answering that benchmarking question, is more accurate than the original model θ, the “inner loop” returns a positive reward signal. If the accuracies are the same, it returns no reward; if θ’ proved to be worse based on the benchmarking question, it returns a negative reward. Now, this process simply repeats with a classic example of Reinforcement Learning, where good SEs are “rewarded” with a positive reward and bad SEs are discouraged with the opposite; through many iterations of this training, SEAL gets good at optimizing the decoder through using the self-edits. One important point to observe is that the SEAL network is updated solely based on the reward signal from the “inner loop”, signaling how well the θ’ model performed relative to θ. separate more Inventing new model frameworks is an arduous task, mostly because extreme caution needs to be taken to ensure that the learning is not corrupted by inherent knowledge or missteps in “signaling” between the loops. The researchers carefully skirted these risks by using decoder-only transformer models that had not been trained on the benchmarking tests they used, meaning that the training evaluations were the first times they had encountered each problem, in turn eliminating the possibility that the model simply “learned the test”. In addition, the model made sure that the evaluations on θ’ were completely independent from that on θ and that the original model never changed across iterations, ensuring that each time SEAL performed SFT to crease a new instance of θ’, it would be based on the exact same θ. not The results were striking; in one particular benchmarking test conducted by the researchers, the model scored a 72.5% success rate, up from 0% without SEAL fine-tuning, demonstrating the insane potential of their framework. If refined and holistically integrated, this framework may become a new industry standard in enhancing AI performance in specific fields or in general. This article is brought to you by Our AI, a student-founded and student-led AI Ethics organization seeking to diversify perspectives in AI beyond what is typically discussed in modern media. If you enjoyed this article, please check out our monthly publications and exclusive articles at https://www.our-ai.org/ai-nexus/read! This article is brought to you by Our AI, a student-founded and student-led AI Ethics organization seeking to diversify perspectives in AI beyond what is typically discussed in modern media. If you enjoyed this article, please check out our monthly publications and exclusive articles at https://www.our-ai.org/ai-nexus/read! https://www.our-ai.org/ai-nexus/read To Learn, or Not to Learn? Regardless of how technically impressive the research team’s achievement is, the far-reaching societal and philosophical implications of this discovery cannot be overstated. I’ve always been a staunch critic of biological computing initiatives (see: Epiphany from the May edition of the AI Nexus Magazine) because I believe that neuronal clusters, like those used in biological computers, are subject to the natural laws because they currently possess the capability for consciousness, and, even if they don’t, are likely to be able to naturally evolve it as a result of plasticity. SEAL is therefore significant beyond a method of improving model performance on benchmarking tasks; it is the first established AI training framework in which an AI model has successfully demonstrated the capability of directly training another AI. Not only does this suggest that we may well be on the path to eventual self-replicating AI paving the way for the AGI singularity, it begs the moral question of whether AI capable of evolving in this fashion should be considered in the context of the rights that we implicitly attribute to living beings like humans and animals. Epiphany May edition of the AI Nexus Magazine May edition of the AI Nexus Magazine AI Nexus Magazine There is a distinction to be made with adaptability and consciousness. We find it permissible to step on a blade of grass since we know that, although it will likely suffer damage, it is not experiencing the animalistic notion of pain since it does not have nerves. However, grass blades are alive, and they demonstrate an uncanny ability to adapt to its surroundings by planting itself in the crevices of concrete slabs. We would, however, hesitate to torture an animal, and I contend this is likely because we are inherently cognizant that feeling pain elicits a much more noticeable response—whimpering or crying, perhaps—which humans, being animals themselves with similar responses to pain, sympathize with. Animals developed pain—a reminder of the fact that they are alive and deserving of some basic rights—over a few millennia of natural evolution, yet I fail to notice a significant disparity between the basal nature of artificial and biological evolution; AI models can, arguably, “evolve” similar processes as pain, and mimic human responses so well that a human, over text or even voice, could not reliably distinguish whether it was an AI or a human who produced them. In fact, this is already happening in the form of a randomized three-party Turing Test, in which AI models like ChatGPT-4.5 have successfully convinced a human interrogator that it was human in over 70% of cases. are have successfully convinced a human interrogator If an AI model acts like a human in every aspect, could it ever be considered a human? Will the trend of AI evolution produce such unique and situationally sensitive models that they start approaching the empirical limit of being “artificial”? Only time can tell. Written by Thomas Yin Written by Thomas Yin

Groundbreaking MIT Research Indicates That AI Can in Fact Teach Other AI Models

MIT-navorsing dui daarop dat AI eintlik ander AI-modelle kan leer

MIT এর গবেষণায় বলা হয়েছে যে আইটি বাস্তবে অন্যান্য আইটি মডেল শেখাতে পারে।

Průlomový výzkum MIT ukazuje, že AI může ve skutečnosti učit jiné modely AI

MIT-Forschung zeigt, dass KI in der Tat andere AI-Modelle lehren kann

Investigación del MIT revela que la IA puede enseñar otros modelos de IA

MITの研究によると、AIは事実上他のAIモデルを教えることができます。

MIT-ийн судалгаа нь AI-ийн бусад AI-ийн загварыг суралцах боломжтой гэж үздэг

Penyelidikan MIT menunjukkan bahawa AI sebenarnya boleh mengajar model AI lain

Hulumtimet e MIT tregojnë se AI mund të mësojë modele të tjera të AI

การวิจัย MIT แสดงให้เห็นว่า AI ในความเป็นจริงสามารถสอนรูปแบบ AI อื่น ๆ

MIT Research o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z o‘z

Ukuhlolwa kwe-MIT Ukubonisa ukuthi i-AI ingathola imodeli ezingaphezu kwe-AI

Stories, News & Mentions about MEDIPHANY

Categories

Trending Topics

Stories, News & Mentions about MEDIPHANY

Mediphany’s Stories on HackerNoon

Categories

Trending Topics