Uri Levine’s new book is a reminder for founders to put the problem first

“Simple can be harder than complex. You have to work hard to get your thinking clean to make it simple. But it’s worth it in the end because once you get there, you can move mountains”

Steve Jobs (from Chapter 8)

I got to meet Uri Levine, the founder of Waze and Moovit, through his involvement in the Zell Entrepreneurship program at Reichman University. Uri is a mentor in the program (where also my co-founder at Remagine Ventures and I met) and apart from his impressive track record as an entrepreneur ($1.4 billion acquisition by Google for Waze and $900M Moovit acquisition by Intel) Uri has also been involved in various startups, especially on the consumer front, as an angel investor.

So needless to say, I was very curious to read his new book “Fall in Love with the Problem, Not the Solution” and indeed, he didn’t disappoint. According to Levine, the cornerstone of a successful business is a profound comprehension of the problem that needs solving, followed by crafting a solution that caters to the unique needs of your target audience. The book is a practical guide for founders on what it takes to build a successful business. It can be read from cover to cover, or ad-hoc based on the need of the founder, for example, how to raise money for investors, firing and hiring and how to scale to a billion users.

It starts from how founders pitch their startups. A solution-focused story starts with “My company does…”. A problem focused story starts with “We solve the ____ problem” or “We help ____ do _____ faster/cheaper/better”.

What’s the gist?

Levine cautions entrepreneurs against becoming overly enamoured with their solution. While the thrill of creating a new product or service can be intoxicating, it’s crucial to remember that the solution is not the ultimate objective. The true goal is to address a problem, with the solution serving as a conduit to that end. The customer should be at the forefront of the entrepreneurship process, not the features or tech (something I talked about in my post on AI startups, “Are you telling a story or pitching features?“).

Levine underscores the significance of customer feedback. He advocates for entrepreneurs to maintain an ongoing dialogue with their customers, soliciting their thoughts on how to enhance their product or service. By actively listening to their customers, entrepreneurs can ensure they’re tackling the right problem and that their solution is fulfilling the needs of their target audience. This is a crucial point that widely known, but often forgotten. Talking to customers BEFORE embarking on a solution is key to understanding the problem, but it doesn’t end there.

Another solid principle Levine advocates is Be willing to iterate and improve your solution based on feedback. No product or service is perfect when it first launches. It is important to be willing to iterate and improve your solution based on feedback from your customers. This will help you to create a product or service that is truly valuable to your target audience. Failure is of course part of the game, but Levine advocates that the only real failure is giving up.

Educational Examples:

While developing Waze, Levine initiated conversations with people about their experiences with traffic. He discovered that people were exasperated by the absence of real-time traffic information, often resulting in tardiness for work or appointments due to traffic congestion. This insight into the problem inspired Levine to create Waze, a real-time traffic navigation app that has since assisted millions of people in avoiding traffic and reaching their destinations punctually.

Similarly, while creating Moovit, Levine engaged in discussions with people about their experiences with public transportation. He found that people were frustrated with the scarcity of information about public transportation schedules and routes, often leading to long waits for buses or trains. This understanding of the problem led Levine to develop Moovit, a public transportation app that provides real-time information about schedules, routes, and arrival times.

These instances illustrate how Levine’s insights have been instrumental in his success in building two thriving companies. If you’re an entrepreneur, I highly recommend reading “Fall in Love with the Problem, Not the Solution”. It’s a valuable resource that can significantly enhance your chances of success.

Key Takeaways for founders:

  • Begin with a profound understanding of the problem you’re attempting to solve.
  • Avoid becoming too attached to your solution.
  • Maintain constant communication with your customers, seeking their input.
  • Be open to refining and enhancing your solution based on feedback.
  • Embrace failure as a part of the process.

If you are an entrepreneur, I encourage you to read Fall in Love with the Problem, Not the Solution. It is a valuable resource that can help you improve your chances of success, especially for first time founders in the early stages. Levine’s insights, derived from his personal journey of building two successful companies, are very action oriented and every chapter ends with ‘Startips’. For example, a reminder that work-life balance doesn’t exist for founders, but if you fall in love with the problem, you will not want (or be able) to do anything else!

Benchmarking LLMs performance


The battle between the various large language models (LLMs) is heating up. I previously wrote about the big tech titans battling the area of AI supremacy, mainly Google vs. Microsoft (via OpenAI) and also the past, present and future of tech wars.

The market share of chatbots, and LLMs (that will later on be sold to enterprises) will be heavily influenced by user perception. That perception is being cemented today is a function of three main characteristics:

  1. Distribution – who can access more users faster
  2. Quality – which product provides better results
  3. User experience – everything around usability, including down time, hallucinations and workflow integrations

Microsoft vs. Google

So far, it seemed like OpenAI was clearly in the lead and as result, it enabled Microsoft to quickly push ahead with AI features in Office 365, Teams and Bing Search. ChatGPT became the fastest growing consumer product and new functionality rolled out in short succession.

However, that narrative seemed to have taken a bit of a dent following Google I/O 2023 last week. Google is throwing everything at AI and positioned the company to make AI its top priority. Google announced it is releasing Palm-2, an LLM to rival GPT. It’s multilingual (100 languages), can do math and reasoning and it is able to code.

Google also opened up BARD, Google’s chatbot assistant and ChatGPT competitor, to everyone (apart from a number of limited countries where Google cannot comply with its regulation). One of the key differences between BARD and ChatGPT is that Bard is connected to the Internet and can provide real time results. ChatGPT is an incredibly powerful tool, but its trained on data up to September 2021 and is only now starting to connect to the Internet via Plugins (the feature is being rolled out slowly only to Pro subscribers).

ChatGPT using GPT-4
Google’s BARD

Google didn’t stop there. The company announced is also releasing new AI tools in Gmail. Google Workspace (docs, spreadsheets, etc), Google photos (to enhance images with AI, detect images that were created by AI in image search), and perhaps most importantly, revamping the search experience and enriching the results from its classic 10 blue links to an AI-powered experience called AI-snapshots.

AI Snapshots by Google Search

Open Source vs. Closed LLMs

Google and OpenAI are the two leading players in the LLM space. Google’s LaMDA is a factual language model that can generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way. Microsoft’s Turing NLG is a generative language model that can generate text, translate languages, and write different kinds of creative content.

But the battle between LLMs isn’t just between the tech giants. Open-source LLMs, such as Meta’s LLaMa, MosaicML, Vicuna-13B (an open-source alternative to GPT-4 which reportedly achieves 90% of ChatGPT’s quality) or RedPajama (which released a 1.2 trillion token dataset that follows the LLaMA recipe) are gaining popularity due to their flexibility and affordability. But asAI researcher Andrej Karpathy has identified several challenges that the open-source LLM ecosystem still faces, including the high costs of pre-training base models.

I asked BARD to compare open source vs. closed LLMs and while it looks like Open-source LLMs are the clear winner it’s not entirely straightforward.

Open vs. Closed LLMs

Closed-source LLMs have several advantages over open-source LLMs. They are typically trained on larger datasets, which allows them to achieve better performance. They are also more tightly integrated with their respective platforms, which makes them easier to use. However, closed-source LLMs are not as transparent as open-source LLMs. This makes it difficult to understand how they work and to identify potential biases.

Open-source LLMs are typically trained on smaller datasets, which limits their performance. They are also less tightly integrated with their respective platforms, which makes them more difficult to use. However, open-source LLMs are more transparent than closed-source LLMs. This makes it easier to understand how they work and to identify potential biases.

The Pepsi Test for Benchmarking LLMs – ChatBot Arena

Chatbot Arena is an underrated tool in the space of LLM benchmarking – A ‘Pepsi taste challenge’ for comparing LLMs.

How does it work?
Users put the same prompt into two boxes and it feeds it to random LLMs (you don’t know which). You get your output back and you rate which one is best in a blind test.

Chatbot Arena LLM comparison

The result is a leaderboard of LLMs which reveals surprising outcomes. For example, the new Claude model from Anthropic is indistinguishable from GPT-4. Take a look at the latest leaderboard table (still doesn’t include Google’s PALM-2).

LLMs leadeboard in ChatBot Arena

High competition and no moat is giving investors jitters

According to the State of AI Q1’23 report by CB Insights, during Q1 2023, AI startups raised $5.4B, a 66% drop from the previous year’s figure and a 43% decline quarter-over-quarter. This was largely unexpected given the supposed favourable market conditions for AI startups.

My belief is that voices are growing louder about the potential bubble in AI where most players, including Google, admit to having no moat.

AI investments down in Q1 2023

Another potential reason to the decline in AI investments is the rising prices, especially in a market that has seen most startup valuations (especially later stages), decline in price.

Prices for generative AI startups continue to climb

As the number of generative AI startups grows exponentially, partly due to the increase in models and APIs made available for the application layer of generative AI, established categories like copywriting for marketing or document summarisation are getting more crowded. Many of the companies competing in these crowded markets will have to pivot or shut down if the cannot become market leaders.

With that said, I believe there’s still a huge opportunity for generative AI startups operating in niches. Rather than rely on a single API, these startups would be wise to combine a number of models and adjust their products to the workflow preferred by their users – connecting to existing tools and adding automation to the way things are done today.

LLMs are here to stay

The battle between the various LLMs is likely to continue for several years. It is too early to say which approach will ultimately be more successful, but there’s a lot riding on it. User perceptions are being cemented now as LLMs continue to expand and power more use cases.

Regardless of which model turns out to be superior, it is clear that LLMs are a powerful new technology with the potential to revolutionise the way we interact with computers. At Remagine Ventures, we are following this space, as well as the application layer of generative AI, very closely and looking to continue investing in cutting-edge teams.

Moral and Ethical Concerns for Generative AI (guest post)

As part of my work at Remagine Ventures, I am constantly impressed by the incredible advancements made possible by generative artificial intelligence (AI) technologies. From natural language processing to image and gaming, the possibilities are endless. We are all dealing here with exciting, scary, and inspiring technologies that will probably drastically change the way we live. I believe that even the greatest minds who are leading this revolution are not quite sure what to expect from it. 

We can probably try to imagine what the next five years to come may look like, but can we think about how the world will look like in 10, 15, 70 years? Those times will come eventually, and while we can’t predict the future, we can definitely start taking some more issues under consideration while embracing those changes into our lives. Thus, as these technologies continue to evolve, it is essential that we consider the moral implications of their use.

Creators in an AI generated content world

For starters, let’s think about all the creators out there: Generative AI technologies, in essence, create new content that did not previously exist. While this can be exciting and innovative, it also raises important questions about who owns the content created by AI. For example, if a machine learning algorithm generates a new song or piece of artwork, who should be credited for that work? Should it be the AI creator, the person who trained the algorithm, or the end-user who requested the output? That’s a big question I am not sure we have a good answer for.

Removing bias from AI training

Another important concern is addressing the potential for generative AI to perpetuate harmful biases and stereotypes. These technologies learn from the data they are trained on, and if that data is biased, then the AI’s outputs will reflect that bias. For instance, if an AI is trained on data that reflects gender or racial biases, it could lead to biased outcomes in the content it generates. This could have far-reaching consequences in areas such as hiring or lending decisions, where automated processes are increasingly being used.

Privacy and the risk of impersonation

Privacy is also a key concern when it comes to generative AI technologies. These tools can create incredibly realistic and convincing content, which could be used for nefarious purposes such as creating fake videos or images that could be used for identity theft or fraud. Also, have you ever thought about what happens with all the information we share with Chat GPT, for example? Or what can be done with all the information about you that the platform already received? Just this month Samsung banned its employees from using ChatGPT after a sensitive code leak.

AI and human creativity

Furthermore, the use of generative AI technologies raises questions about the role of humans in the creative process. As machines become increasingly capable of creating original content, it is natural to wonder whether humans will be needed at all. While it is unlikely that AI will completely replace human creativity, there is a chance that these technologies could change the value placed on human creative work, which could have negative consequences for the arts and culture as a whole. With creative platforms such as Midjourney that are evolving at an exponential rate, there are a lot of questions to be asked about how we would see and appreciate creations in the future. I believe that creativity is here to stay – but the way we define it, perceive it, and evaluate it will change drastically in the near future.

To address these moral concerns, researchers have proposed methods to de-bias data and create new legal frameworks for AI-generated content. Additionally, it is important to ensure that AI is used in ways that benefit society as a whole, while also minimizing its negative effects. At Remagine Ventures, we are committed to supporting the development of AI technologies that are ethical, responsible, and respectful of human values.

As we continue to explore the potential of generative AI technologies, it is important that we take these moral concerns seriously. By doing so, we can ensure that these tools are used in ways that benefit society as a whole, while also protecting the rights and privacy of individuals. However, we should also remember that concerns are a normal part of every major technological revolution, and that it is our responsibility as investors, creators, entrepreneurs, and users to navigate this new era of opportunities while remaining mindful of ethical considerations. 

Generative AI technologies have the potential to transform the way we live and work, but it is essential that we consider the moral implications of their use. By taking a proactive and responsible approach to AI, we can create a better future for everyone.

Unveiling Possibilities: Learnings from TED 2023 in Vancouver

I got back from the picturesque city of Vancouver a couple of weeks ago, after attending TED 2023 (the main conference) for the second time. Apart from being one of the best events from a production point of view (lots of attention to detail from communication to infrastructure and design) it was an opportunity to learn, network and get inspired. The theme of this year’s conference was ‘Possibility’ and there was a significant focus on AI and generative AI, and the possibilities they unlock.

It’s hard to summarise a whole week in a short blog post, but I wanted to share a few thoughts on key themes from the conference and what I learned.

Theme 1: The Power of Generative AI

The first AI talk was by Greg Brockman, co-founder and CEO of OpenAI. It was a glimpse and live demo of some of the features coming to ChatGPT in the next few months including the plugins integration, that connects ChatGPT to the Internet, data visualisation and graph creation with code interpreter and more. I recommend you watch it (video of the talk embedded below).

Greg Brockman, co-founder and CEO of OpenAI (Ted talk April 2023)

Perhaps more interesting was the Q&A that followed with Chris Anderson, where Greg was asked why did they unleash such a powerful tool on the world without taking into account the many implications it will have on society and also should AI development be slowed down?

Greg’s response (and Sam Altman’s as well in the video below) is a window to OpenAI’s philosophy on this. In a nutshell, they believe that AI development is inevitable, and that it cannot be locked in a lab to be developed in secret. They believe that having the product in the hands of users where the ‘rubber hits the road’ provides them with valuable feedback on how the products should be improved.

Theme 2: AI in Science

Karen Bakker, a professor at the University of British Columbia and conservation technology researcher, gave a talk on how AI is enabling us to capture the world of ultra or infra sonic sounds, which animals use to communicate with one another, and start to decipher what some of those sounds might mean. It was a thought provoking session as it might be possible to achieve interspecies communication as we continue to decode how Orcas, dolphins, bats, etc talk to one another. Using generative AI, scientists were able to reproduce these sounds to try to communicate with bees, for example, but so far with little success.

Scientists are even able to translate some variations of animal speech, while generative AI is able to imitate some of these sounds, allowing us to communicate with nature like never before – and bringing along some difficult challenges, too.

The TED blog

Can you imagine a Dolphin giving a TED talk about the impact of warming oceans in a few years? Karen goes deeper on this topic in her book, The Sounds of Life.

Theme 3: AI in Education

When it comes to ChatGPT and generative AI, you can be certain that students of all ages are early adopters, as it makes it much easier for them to turn in their homework ;-)

TED featured two interesting examples of AI in education that showed the potential of generative AI to give every teacher an AI teaching assistant and a personal tutor for every student. The first one was Sal Khan, founder of Khan Academy. Sal demonstrated Khanmigo, a GPT-4 powered teaching assistant.

The second was a talk by the Luis von Ahn, CEO of DuoLingo, the most popular language learning school, which started offering a new AI-powered language tutor to consumers, powered by GPT-4. The interesting part was their AI-powered gamification to motivate users to learn. For example, DuoLingo figured out that the best time to send users notifications needs to be personalised. They are most likely to be available for a session at the same time they opened the app in their previous session. In addition, they communicate with users in a very natural language and increase engagement as a result. Here’s an interesting article on what DuoLingo does differently to increase retention.

Theme 4: the risks and limitations of AI

Overall, TED did a good job balancing the promise of AI with the dangers it poses and its limitations. Computer scientist Yejin Choi spoke about the importance of giving AI human values. In her view, many AI systems brings three big issues with it:

  1. AI models are expensive to train,
  2. Their power is concentrated to only a few tech companies and
  3. The environmental impact is massive

Gary Marcus spoke about the risk AI poses on spreading disinformation and called for governments to regulate this technology before it’s too late. Perhaps the most controversial talk was by Eliezer Yudkowsky, who was invited to speak only a week before the conference and gave a 6 minutes talk that he read from his phone. His message was a scary one: superintelligent AI could probably kill us all and by the time we realise it, it will be too late. Eliezer recently recorded a long podcast episode with Lex Fridman if you want to learn more about his views. Also, Tom Graham, the founder and CEO of Metaphysic, the company behind deep fake Tom Cruise, talked about new opportunities (and risks) for this technology in media and entertainment. Other risks that were discussed: AI weapons, fake news, copyright infringement, bad actors leveraging AI agents… and so on. No real answers to these risks, apart from a sense that the industry should agree on an ethical code for AI and help regulators set some boundaries.

Final thoughts

There was much more than AI discussed, and I would be remiss not to mention the work of some inspiring non-profits that presented, talking about issues of social justice, racial inequality, climate or the member of Pussy Riot that served two years in prison and is wanted by Russia. There were also the classic motivational speakers that sometimes made me move uncomfortably in the chair, like I’m in the SNL version of TED, but generally speaking I was more attracted to the talks that had a real message and story to tell, vs. great delivery/theatrics. Some of the most impactful talks had nothing to do with AI – a death dula that accompanied hundreds of people and their families to help them prepare for their death, is possibly one of my favourites and made me think about my own mortality. Another talk that landed was about the importance of supporting boys (not at the expense of girls, but the balance swung too far the other way). I encourage you to watch them all on Youtube when they get published.

Overall, the talks were only part of the experience. Much of my learning came from conversations with interesting people, including Reed Hastings, founder of Netflix, Yat Siu, co-founder and chairman of Animoca, and many other professors, entrepreneurs, VCs and philanthropists. I ran into friends, ex colleagues from Google and made new friends.The TED community is what genuinely makes TED special.

TED made me put our work at Remagine Ventures in context, and realise that despite all the shit we hear in the news, there’s a lot of progress being made that doesn’t normally get the spotlight. It was also a good reminder that the information we get used to consume, and what you will read in the news is not necessarily what’s most important, so stay curious and keep on learning. We are living through a time of unprecedented progress in science and technology, but despite all that tech (or maybe because of it) there’s no replacement to human connection and interaction. Thank you for reading.

The Evolution of Tech Wars: Past, Present, and Future

Like many ideas, this post started with a Tweet…

Having spent the past 25+ years in tech, I’ve lived through a number of tech “wars”. In the early days of PCs, it was the semiconductor wars. In the early days of the Internet, we witnessed the browser wars and as technology continued to advance, we saw the number of wars, and the number of players involved in them grow. So I decided to nerd out on this topic a bit and expand the list or tech wars that were, and those that are yet to fully play out.

In this post, I’ll delve into some of the most notable tech wars, both past and present, and explore the battles yet to come, including several new additions to the list.

Past Tech Wars

  1. Smartphone Wars: Apple took the lead in the smartphone wars, capturing a significant share of the market with the iPhone. Android may have a larger global market share, but Apple’s loyal customer base, strong brand, and profitable ecosystem made it the clear winner.
  2. Console Wars: While Sony has a bigger market share in console sales, Microsoft emerged victorious in the console wars with its Xbox line, overcoming competition from Sony’s PlayStation and Nintendo’s gaming systems. A strong focus on online services, exclusive games, and innovation secured Microsoft’s position at the top.
  3. App Store Wars: Apple dominated the app store wars, building a walled garden around its App Store and generating significant revenue from app sales and in-app purchases. Google’s Play Store remains a strong contender, but Apple’s control over the iOS ecosystem has proven unbeatable.
  4. Bundle Wars: Apple won the bundle wars with the launch of Apple One, a suite of services including Apple Music, Apple TV+, Apple Arcade, and iCloud storage at an attractive price. By consolidating its services, Apple locked users into its ecosystem and claimed victory.
  5. Semiconductor Wars: In the early days of the PC, companies like Intel, AMD, and Motorola competed to create the most powerful and efficient microprocessors, driving the evolution of personal computing.
  6. Browser Wars: In the early days of the internet, the browser wars saw fierce competition between Microsoft’s Internet Explorer, Netscape Navigator, and later, Mozilla Firefox and Google Chrome. The war shaped the way users interacted with the web.
  7. Search Engine Wars: In the late 90s and early 2000s, search engines like Yahoo, AltaVista, and Lycos competed to provide the best search experience, until Google ultimately emerged as the dominant player.

Current Tech Wars

  1. Streaming Wars: The streaming wars are still raging, with platforms like Netflix, Amazon Prime Video, Disney+, and Apple TV+ competing for viewership and market share. The future remains uncertain as each player invests in content and innovation to attract and retain subscribers. I wrote about the streaming wars on VC Cafe back in December 2019, and it’s interesting to see how many of my predictions came true: sports, gaming and live content to name a few.
  2. Cloud Wars: As businesses increasingly adopt cloud services, the battle between cloud providers like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP) intensifies. Each company is vying for dominance, offering unique features and pricing models to attract customers. We are seeing some of this play out in the generative AI space, where the cloud providers are active investors to attract future revenue. I wrote about this in a recent post on VC Cafe.
  3. Wallet Wars: Digital wallets are transforming the payment landscape, leading to a fierce competition between companies like Apple Pay, Google Pay, and Samsung Pay. With secure, convenient, and contactless payments, the wallet wars are set to redefine the future of transactions.
  4. Blockchain Wars: As the potential of blockchain technology becomes increasingly apparent, tech giants and startups alike are racing to develop and implement solutions across various industries. Ethereum, Binance Smart Chain, and Cardano are among the leading platforms in this ongoing battle for blockchain supremacy.
  5. AR/VR and Metaverse: The race to create the ultimate augmented reality (AR), virtual reality (VR), and metaverse experiences is heating up. Tech giants like Facebook (Meta), Apple, and Google are investing heavily in AR and VR technologies, competing to define the future of immersive experiences. My post on the 10 companies investing billions in the metaverse is perhaps a bit dated now as the market changed, but the rumoured new Apple XR device might flare things up again.
  6. Cloud Cybersecurity: With the rise in cloud adoption, securing cloud infrastructure has become a critical concern. Companies like Palo Alto Networks, CrowdStrike, and Zscaler are vying for supremacy in the cloud cybersecurity market, offering cutting-edge solutions to protect valuable data and systems.
  7. US-China Chip Wars: As the global semiconductor industry faces supply chain challenges and increasing demand, the rivalry between the US and China intensifies, with both nations striving to establish dominance and self-sufficiency in chip manufacturing. This one is not only a tech war, but also geo-political and involves national security interests.

Upcoming Tech Wars

  1. Quantum Computing: As traditional computing approaches its limits, quantum computing emerges as the next frontier. Companies like IBM, Google, and Intel are competing to develop powerful quantum computers capable of solving complex problems beyond the reach of classical computers.
  2. Smart Car Wars: With electric and autonomous vehicles gaining traction, the smart car wars are set to intensify. Tesla, Ford, and GM face competition from tech giants like Apple and Google, who are investing heavily in self
  3. Quantum Computing Wars: As classical computing approaches its limits, the race for quantum computing supremacy heats up. Tech giants like IBM, Google, and Intel are competing to develop powerful quantum computers that can solve complex problems beyond the reach of traditional computers, with potential applications in cryptography, drug discovery, and optimization.:
  4. Large Language Model (LLM) Wars: AI advancements have led to a competition between tech giants like OpenAI, Google, and Microsoft in the large language model (LLM) space. These AI models, capable of understanding and generating human-like text, will transform industries and unlock new possibilities. New contenders including HuggingFace open source model, Anthropic, AI21 Labs Jurassic, Meta’s LLaMa and others continue to keep this space red hot.
  5. Space Race 2.0: Private companies like SpaceX, Blue Origin, and Virgin Galactic are competing to revolutionize space travel and exploration, with goals such as commercial space tourism and Mars colonization.
  6. Biotech Wars: The race to develop cutting-edge biotechnologies is accelerating, with companies vying to create breakthroughs in gene editing, personalised medicine, and synthetic biology.
  7. Robotics Wars: As robots become more sophisticated and capable, the competition among robotics companies like Boston Dynamics, SoftBank Robotics, and ABB intensifies, with a focus on developing robots for various industries and applications.
  8. Green Energy Wars: With the global push towards renewable energy, companies like Tesla, Vestas, and First Solar are battling to dominate the market for solar, wind, and other clean energy technologies.

Where does this leave us?

The tech industry has always been a battleground for innovation, market share, and consumer loyalty. As we continue to witness the unfolding of existing tech wars and anticipate the emergence of new ones, one thing is certain: the race for supremacy will drive technological advancements that transform our world in ways we can only imagine.

As a venture capitalist, these tech wars offer both opportunities and challenges. Investing in competitive markets demands a deep understanding of the landscape, the players involved, and the potential for market disruption. Peter Thiel, in his book “Zero to One,” emphasizes the importance of backing companies that aim for monopoly, stating, “Competition is for losers.” The quote highlights the significance of investing in businesses that create and dominate new markets, rather than merely competing in existing ones.

When placing bets in competitive markets, it’s crucial to identify the potential game-changers – companies that redefine industries, create new markets, and ultimately emerge as monopolies. As the tech wars continue to unfold, we must remain vigilant, adaptable, and ready to seize the opportunities that arise from the ever-changing landscape.

AutoGPT is Generative AI goal seeking

ChatGPT has barely been out for six months, but it seems like generative AI automation has just taken a leap forward with AutoGPT.

Rather than produce a single answer based on a single prompt, AutoGPT is capable of breaking down a task to a number of sub tasks, and continue to perform actions autonomously until the goal is reached. Driven by GPT4, AutoGPT is able to code, execute scripts and push the boundaries of what is possible with AI. In this post we’ll look at what is AutoGPT, how can it be useful and the risks and limitations of the technology.

Auto-GPT is an experimental open-source application showcasing the capabilities of the GPT-4 language model. This program, driven by GPT-4, chains together LLM “thoughts”, to autonomously achieve whatever goal you set. As one of the first examples of GPT-4 running fully autonomously, Auto-GPT pushes the boundaries of what is possible with AI.

Github

AutoGPT is essentially goal-seek

Auto-GPT is an open-source application, created by game developer Toran Bruce Richards. It can help reduce decision-making errors by eliminating bad data points and making it more robust to change.

The user gives the AI a goal, and it will try to accomplish the goal on its own, by breaking it down to a series of actions. It’s To get a sense of how it works you can try this demo on AgentGPT. While it doesn’t yet complete the tasks, you can see the steps the AI would take to get there.

Take a look at this demo video:

How does AutoGPT work?

Auto-GPT is publicly available on GitHub but to use AutoGPT today you’d have to be a developer and paying user of OpenAI, as it requires a OpenAI API key. Once you have your API key, enter it in the settings of AgentGPT, which is the web interface for AutoGPT. If you’re up for the task, here’s a step by step guide on installing and implementing AutoGPT.

Auto-GPT is an unsupervised learning AI tool that uses OpenAI’s GPT-4 and ChatGPT application programming interface (API) models, with pricing based on token usage, making it suitable for tasks such as content generation and coding projects.

Cointelegraph

AutoGPT essentially creates its own list of prompts.

These agents are self-contained systems that use modern generative AI models to automate tasks. Most agents use OpenAI’s ChatGPT and GPT-4 as a base, but several other homespun agents also take in generative AI image and voice models to create some surprising, if sometimes creepy results. These systems feed the AI’s outputs back into themselves, creating a program that can run semi-autonomously with an overarching goal

Gizmodo

Examples of AutoGPT implementations

AutoGPT is merely one month old, but there’s been a number of use cases shared on social media that showcase its potential.

1. Manage finances and save money

2. Conduct research

3. Complete the tasks on your Todo list

4. Build a website from scratch

5. Write code and automatically debug it

6. Read recent events and prepare a podcast outline

And the list goes on and on.

What could possibly go wrong?

I remember a cybersecurity talk, where one of the ‘good guys’ working on behalf of organisations to defend against attacks said “a hacker needs to find one vulnerability to be successful. We need to find all of them”. As this technology scales, you can imagine AI agents working tirelessly to execute phishing attacks, spread disinformation and cause all sorts of havoc, like in the case of ChaosGPT, an AI tool that seeks to destroy humanity (I’m not kidding!).

Then, there’s the accuracy and hallucinations. As shared by TechCrunch, depending on what objective the tool’s provided, Auto-GPT can behave in very unexpected ways.

One Reddit user claims that, given a budget of $100 to spend within a server instance, Auto-GPT made a wiki page on cats, exploited a flaw in the instance to gain admin-level access and took over the Python environment in which it was running — and then “killed” itself.

Techcrunch

Like GPT-4, anything that’s built with it is prone to inaccuracies and hallucinations. Left unchecked (which is the whole point of AutoGPT), users can end up with very wrong information, articulated convincingly.

As described by Clara Shih, the CEO of Salesforce’s Service Cloud and an Auto-GPT enthusiast:

“Auto-GPT illustrates the power and unknown risks of generative AI. For enterprises, it is especially important to include a human in the loop approach when developing and using generative AI technologies like Auto-GPT.”

AI agents will continue to be developed to bring automation to tasks for both consumers and companies. Other projects in the space of AutoGPT include BabyAGI and Genesis. As the co-founder of OpenAI tweeted, AutoGPT might be the next phase of prompt engineering

What can online communities learn from Eve Online?

Online gaming has become a popular form of entertainment over the past few decades. One game that has stood out as a leader in the industry is EVE Online. This massively multiplayer online game (MMO) has been around since 2003 and has built a dedicated following of players who have formed their own unique online communities within the game.

I recently read the book “Empires of EVE: A History of the Great Wars of EVE Online” (by Andrew Groen), which is a book that delves into the complex political landscape of EVE Online, highlighting the various alliances and coalitions that have shaped the game’s history. For those unfamiliar with the game, at times, 50-60k players would battle each other in one massive fight, coordinating tens of thousands of people simultaneously.

In case you’re not familiar with the game, watch this short clip:

Lessons for running large online communities

While the book is primarily focused on the game’s wars and conflicts, it also provides valuable insights into how to build and maintain effective online communities. I was impressed with the coordination, teamwork, huge effort and leadership displayed by different alliances. Some of the efforts involved spies, lots of disinformation and psych warfare tactics on message boards and heists. Not kidding. More on this below…

But, as mentioned above, to me one of the key takeaways from the book is the importance of strong leadership in online communities. The successful alliances in EVE Online were able to thrive because they had strong, dedicated leaders who were able to coordinate and motivate their members. These leaders were able to create a sense of unity and purpose within their alliances, which helped them to achieve their goals and weather difficult times. A lot of alliances broke down because players felt it became a lost cause. The successful ones did not. 

One such example is “The Mittani” , one of the most well-known players in EVE Online. He is the founder and CEO of Goonswarm Federation, one of the most successful alliances in the game’s history. In-game, The Mittani is known for his cunning and strategic prowess, as well as his ability to rally his troops to victory. He is also known for his charisma and ability to engage with his fellow players. 

Another key factor in successful alliances was a unique culture that players identified with inside the game and on message boards, with some even meeting in private. Communication also played a vital role in effective alliances, as the most successful ones were able to keep their members informed about important events and decisions, creating a sense of transparency and trust within the alliance. 

The book also highlights the importance of having a clear mission and vision for an online community. The successful alliances in EVE Online were able to achieve their goals because they communicated a clear idea of what they wanted to accomplish to their members. This helped to create a sense of direction within the alliance, which helped to motivate members to work towards their goals. Think OKRs. Not all alliances’ purpose was to wage war or attack, lots of them had more industrial/commercial objectives. 

What can go wrong?

The book also shows many examples of what not to do. Conflict can both bring a community together and tear it apart. And because players took this game very seriously, some of the negative sentiment that started inside the game had real-world consequences. This might be one of the most important lessons here: as we spend more time inside online communities, we bring with us our ego and sensitivities. These can easily transfer back to the real world. In Eve Online’s case, there were even death threats that occurred. 

Another example of antisocial behaviour on Eve Online has been impersonation and theft. Recently, a sophisticated heist resulted in the theft of 2.2 trillion ISK—or $22,300 in real money. It’s not the first time theft occurred on the platform, and serves as a reminder that with all the safeguards in place it’s still challenging to catch cheating, impersonation and social-engineering on these platforms.

Ultimately, EVE Online provides valuable insights into what it takes to build and maintain effective online communities: 

These lessons are applicable not just to online gaming communities, but to any online community that wants to be successful and thrive.

The ever growing Israeli Generative AI landscape April 2023

*Originally published on Calcalist, April 20 2023

Since the day ChatGPT was introduced in November 2022, it has become the fastest growing consumer product, putting the words ‘generative AI’ in the mainstream. More and more, people are experimenting with these new technologies that enable anyone to write better, brainstorm, invent, craft and get inspired with text, images, video, voice, music and more. Every day, we hear about new products and services that are meant to enhance human creativity and productivity. 

Generative AI is advancing at a breakneck speed. The launch of GPT-4 in March 2023, with its 1 trillion parameters and the new plugins that will connect ChatGPT to the Internet. We are seeing Google, Microsoft, Amazon and Meta rushing to integrate generative AI features into their products for consumers, enterprises and developers, and every week we learn of new mega rounds into AI research teams looking to bring generative AI technology to every field, from health to education, commerce, entertainment, etc. 

It’s a global phenomenon, but to be successful in AI, startup hubs require a density of AI engineers, access to data and compute power and access to capital. The US and China are in the lead, but Israel has certainly earned itself a place as one of the leading centers of excellence in the global AI ecosystem with companies like AI21 Labs, which is developing a large language model Jurassic, which can be considered as an alternative to GPT. 

But it’s not only the AI startups that are active in the generative AI space. Many established tech companies like Wix, Lightricks and Fiverr, are embracing generative AI to power-up their products with richer consumer experiences which add new creative capabilities and automation. For example, Wix uses ChatGPT to enable users to automatically write the copy for their new] website, and Lightricks enables users to create AI portraits powered by Stable Diffusion. 

We published the first version of the landscape in February 2023, which included 40 startups. In just a couple of months the current version includes 68 startups. We know a lot of founders in this space choose to remain in stealth and expect the actual number of Israeli startups working on generative AI to be much larger and grow quickly. 

At Remagine Ventures we started investing in the generative AI space very early back in 2019. Several of our limited partners are media and entertainment strategics: broadcasters, publishers, telcos etc, so content creation, distribution and monetisation has always been close to our core focus. While we didn’t yet have the term ‘generative AI’, we called it “synthetic media” or “creative automation” and made 5 investments in the space so far (3 are disclosed on our website) and have been closely following the growing landscape of Israeli generative AI startups. We believe that generative AI will become mainstream when it goes from playful to useful, and generative AI tools will exist for every role in the company across industries. 

While ChatGPT is most famously known, there are over 750 generative AI globally. Some focus on the core model level, developing large language models or diffusion technology for image and video creation, and others operate in the application layer, powering their generative AI experiences by combining one or a number of APIs that are either open source or commercially available. 

Our generative AI portfolio includes HourOne, a text to video solution that enables companies to create high quality presenter-led video from text without the need of a camera. The characters can speak any language and HourOne uses generative AI also to help users come up with the script, based on a text prompt; Munch uses AI to create automated short clips from long form video content leveraging generative AI for descriptions, subtitles and tags. Munch and HourOne are two examples of startups that can dramatically reduce the cost and time it take to create high quality content; Piggy, which is in the application layer of generative AI, enables users to create mobile presentations based on a single prompt, by combining multiple generative AI APs in text, images and design to make content creation on mobile easier. We made two new investments in this space which are still in stealth.

The Israeli AI ecosystem

The global startup ecosystem is going through a rough patch and Israeli startups are no exception. Israeli venture capital investments in Q1 2023 saw a 70% decline compared to the same period last year. It’s a bit odd, considering the fact that innovation was never in higher demand, there’s a growing cohort of talent which is becoming available for a variety of reasons and are willing and able to form new startups and venture capital funds have record dry powder levels, at least on paper.  

The talent pool for AI startups in Israel is deep. It includes not only 8200 alumni, but also ex-engineers at Microsoft, Google, Meta, Intel, Nvidia and others, many of which develop AI technologies, as well graduates and researchers from the Technion, Hebrew University, Weizmann Institute and other leading Israeli academic institutions, which are responsible for cutting edge peer-reviewed research in the generative AI space. This talent pool, combined with Israeli entrepreneurs’ high disregard of the impossible and an active early stage investment ecosystem make me optimistic about the future of this space in Israel. 

We’ll continue updating the landscape on a regular basis. If your company is missing, please scan the QR code to make sure it’s captured for the next version. We love partnering with founders early in their journey and are particularly excited about use cases for generative AI in entertainment, education and gaming as well as infrastructure for generative AI and model training.

The big challenge for Generative AI is GPU capacity and server costs

“The unanticipated demand to create AI software has outstripped the cloud service capacities of Amazon Web Services (AWS), Microsoft, Google Cloud, and Oracle”

The Information

The big challenge in generative AI is GPU shortage and costs. This affects both big players like OpenAI /Stability AI as well as small startups. That’s partly why the big cloud platforms will continue to be active investors in the space. They get their investment dollars back in the cloud bill. The recent $10 billion investment by Microsoft to acquire 51% of OpenAI (much of it in the form of cloud credits) or Google’s $300M investment in Anthropic are good examples of this.

The costs should come down with time, but for now, there’s a mild GPU shortage and the costs are significant. ChatGPT costs approximately $100,000 per day or $3 million per month to run on Microsoft’s Azure Cloud, with each word generated costing $0.0003. Assuming the continued exponential growth of ChatGPT, which has become one of the fastest growing consumer products, the company needs deep pockets to continue to providing the service.

In comparison, Google spends about $100 billion a year on infrastructure costs. It translates roughly to a cost of 5 cents for every query we type into the search box. The difference is that Google knows how to make $2 for every $1 they spend, using Adwords. When it comes to monetisation, OpenAI is expected to make $200M in 2023 and $1Billion in revenue in 2024.

The big winner from this rapid increase in demand is Nvidia, the leader in the market for chips designed to excel at AI computations. Nvidia GPU demand has seen its stock nearly double in the first 3.5 months of 2023. But even Nvidia is two to three months behind on new order fulfilment for cloud server chips. Intel, once the leader in semiconductors, has a lot of catching up to do, especially as the US chips act is trying to bring manufacturing home to the US.

Regardless of cost, it’s clear that generative AI and ChatGPT are rapidly changing consumer behaviour, and the big tech generative AI race is on as previously covered on VC Cafe. Generally speaking, Google has little to worry about when it comes to search marketshare as you can see in the chart below. However small, Bing is growing given their Bing AI integration and Samsung’s consideration of possibly replacing Google search with Bing has put Sundar Pichai in panic mode.

Google’s response in AI so far has been perceived to be more of a follower trying to catch up with BARD and by slowly making AI features available in GSuite, compared with Microsoft’s break neck pace of AI integrations in Bing, Office 365, Microsoft Team’s etc. Google is hoping to change this perception with “Project Magi”, a new search experience would offer users a far more personalised experience than the company’s current service, attempting to anticipate users’ needs as you can see in the video below.

Amazon also positioned itself in the generative AI space with “Project Bedrock” which provides a way to build generative AI-powered apps via pre-trained models from startups including AI21 Labs, Anthropic and Stability AI. Available in a “limited preview,” Bedrock also offers access to Titan FMs (foundation models), a family of models trained in-house by AWS.

The latest entrant to the generative AI space is Elon Musk, who after signing a letter requesting to slow down the training of new AI models is rumoured to be starting a company to rival OpenAI, under the newly purchased X.ai domain. Even at this stage in the game, the opportunity is huge. According to Ori Goshen, co-founder and CEO of AI21 Labs, an Israeli startup developing LLMs that also rival OpenAI:

“Every business will enter the world of generative AI, because every other business in its field will implement solutions from this world“.

For now, the main implication of this for generative startups is that they should either have a clear business model, deep pockets, or a deal with one of the leading cloud providers to be able to serve customers despite limited cloud resources and costs.

Why Generative AI’s impact on entertainment will come fast and hard

 “A tsunami is coming, we can either ride it or get wiped out by it. But it’s going to be really fun to ride it, and it’s going to make us faster and better.”

Nicholas Carlson, Editor in Chief of Insider.com

For the past decade, I have been involved in the media and entertainment industry, both as a worker and an investor. During that time, I’ve witnessed several hype cycles come and go, from AR/VR to the dream of “watch-click-buy” and addressable TV, to the Metaverse, and more. However, the current hype around generative AI feels different. Naturally, a technology like generative AI has the potential to transform and impact many industries, if not all. Many intelligent individuals and consulting firms have written about various industries and their potential use cases for generative AI. To me, however, the entertainment, media, and gaming industry seems the most obvious industry to be disrupted by this technology, far beyond any other. In its simplest form, my argument is the following: 

  1. The entire entertainment industry is driven by a creative process that creates and combines text/code, images, audio & video into compelling storytelling. Whether that is for TV, a game, a newspaper or a podcast. 
  2. What do the current fundamental generative AI models allow you to do? Create text/code, images, audio and video. 
  3. The entertainment industry is one with very low barriers to entry and moats. 

So the elemental tasks that generative AI is currently great at are also the fundamental building blocks of the entertainment industry. The potential to democratise content creation, decrease cost structures and increase productivity are enormous and exciting – and scary at the same time. 

The 4 elements of creation of generative AI: Text, audio, images and video

Generative AI technology is revolutionising the way we create content by providing individuals with the ability to generate text, code, images, audio, and video using machine learning algorithms. At its core, generative AI technology uses transformer models, a process called deep learning, which involves analysing vast amounts of data to identify patterns and generate new content that mimics the characteristics of the original data. This technology has enabled users to create high-quality content without the need for extensive technical expertise, making it accessible to everyone. 

For instance, generative AI can be used to generate realistic images, even of objects or scenes that don’t exist. It can also be used to generate realistic speech, music, or even entire songs. Furthermore, generative AI can generate entire paragraphs of text that sound like they were written by a human, as well as generate code for software development. With the use of generative AI technology, individuals can create content that meets their specific needs and preferences, without the need for specialised skills or resources. Ultimately, this technology is democratising content creation and enabling everyone to express their creativity in new and innovative ways.

These 4 elements of creation are also the building blocks of media & entertainment

At its core, media, entertainment, and gaming are all built from the same 4 building blocks: text, images, audio, and video that are coded together to create immersive and engaging experiences for users. For example, ProSiebenSat.1 Media, a German broadcaster (and one of our investors at Remagine Ventures) is using Hour One, (also a Remagine Ventures portfolio company), to create short news stories from text. What would usually take hours to film in a studio, produce and edit now takes a few minutes – and anyone can do it.

Combining these four elements and using generative AI technology is also what another portfolio company of ours does. Piggy allows anyone to create powerful and informative stories with a single prompt – all on your mobile phone. For example, here’s my piggy on the ‘Best rap albums of the 90s‘. In the world of gaming, the combination of these elements is used to create unique and engaging experiences that challenge players and keep them coming back for more. For example, in role-playing games, text and dialogue are used to create complex and dynamic characters that players can interact with, while images and video are used to create vivid and immersive environments that players can explore. Overall, the combination of text, images, audio, and video is essential to the creation of media, entertainment, and gaming experiences that capture the imagination and engage the senses. 

The entertainment industry is perhaps the most susceptible to disruption among all industries

In recent decades, the emergence of new tools and technologies has made it increasingly easier for users to generate and distribute content, such as blog posts, videos, and simple games, on platforms like Roblox or Minecraft. Moreover, with the advent of the internet, social media, and online marketplaces such as Amazon and Shopify, the distribution channels in the entertainment and media world have broken down. The rise of platforms like YouTube and Twitch has further accelerated this trend. As a result, the barriers to entry that once existed in the entertainment industry have virtually disappeared in a short period of time.

From expensive TV networks to trucks distributing newspapers and magazines, the erosion of moats along the value chain has facilitated the emergence of influencers and thousands of individuals who can now create and distribute content worldwide. In 2016, a survey of executives by the Harvard Business Review (HBR) identified media as the industry most impacted by digital disruption.

When compared to other industries, such as healthcare or fintech, which require more complex products, have entrenched distribution channels, and face greater regulatory scrutiny, it becomes clear why the media and entertainment industry is more susceptible to disruption. 

Despite the relative ease of disruption, the “90%-9%-1%” rule of engagement has long been a guiding principle in the industry. This rule of thumb suggests that 90% of internet users consume content, 9% comment and edit on other people’s content, and only 1% create content. However, with the emergence of user-generated content (UGC) platforms, this rule of thumb has been evolving in recent years. Companies like Minute Media (a Remagine Ventures portfolio company), Roblox, TikTok, Substack, and a host of video creation and editing tools have made it easier for users to create and distribute content. Consequently, more and more users are becoming content creators, shifting the balance of the 90%-9%-1% rule of engagement.

Generative AI technologies are about to take UGC content creation to a whole new level

It’s therefore no surprise that our sector is the biggest user of generative AI technologies. Generative AI can be used to generate high-quality audio and video content (e.g. our investment in HourOne), such as music and movies, by analysing vast amounts of data to create content that is on par with professionally produced content. This technology can also be used to create engaging and informative social media content that captures the attention of users and drives engagement. We have made an investment in this category as well with Munch. 

Generative AI will be used to create realistic and immersive gaming experiences by generating environments, characters, and even entire game mechanics that mimic the style and feel of established games. Our latest investment, not public yet, is aiming to create exactly that, a gaming platform where everyone and anyone can create games via a few prompts. 


Overall, generative AI technologies are enabling a new era of creativity and innovation in the entertainment and content production industries. With the ability to generate high-quality content at scale using generative AI tools, anyone with an idea a computer or a smartphone can become a content creators. Incumbents, including news rooms, are paying attention too and experimenting with generative AI as well. For example, Business Insider and Buzzfeed started using AI to write articles.

As generative AI continues to evolve and improve, we can expect to see a new wave of content creators and innovators emerge, changing the way we create, consume, and interact with media forever. I have written about the long-term impact that generative Ai technology will have on the world of media before. This explosion of synthetic content will create new challenges that need to be solved and we look forward to partnering with exceptional entrepreneurs to tackle them

Thoughts by Kevin Baxpehler, edited by ChatGPT, images created by Midjourney

Exit mobile version