Good morning, and happy Friday.

Elon Musk will soon be essentially removing the block feature on Twitter, something that appears to be in violation of Apple and Google’s app store terms of service.

We’ll see if they do anything about it.

— Ian Krietzberg, Editor-in-Chief, The Deep View

In today’s newsletter:

• ⛴️ AI for Good: Maritime crimes

• 🤖 Tesla unveils Optimus upgrades

• 📊 TSMC spikes on ‘monster’ third quarter

• 🚨 Competition and power consolidation in AI

AI for Good: Maritime crimes 

Overfishing has been a known problem for decades, now. But since it occurs in the middle of the ocean — alongside any other maritime crime imaginable — observation of illegal fishing or illegal maritime activity is nearly impossible, which makes enforcement nearly impossible, which allows the problem to continue. 

The details: A startup founded in 2021 — the Nautical Crimes Investigation Service (NCIS) — aims to bring observability, and, in so doing, enforcement, to maritime crimes. Their tool of choice? Artificial intelligence. 

• The startup began by assembling a thorough database of maritime crimes, linked to the vessels that committed them. Then, it applies tracking technology to monitor signals from ships, enabling the tracking of those vessels. 

• Crossing those movements with an AI tool designed to predict the likelihood of an individual ship committing an environmental crime (because of NCIS’s database), the startup is able to track maritime offenses, connect them to ships and provide that data to non-profits that work with national Coast Guards and Interpol. 

“The goal is to increase transparency, accountability, protection of human rights and due process in crime detection,” NCIS says. 

Build human-like voice agents with Deepgram

Introducing a unified Voice Agent API powered by the industry's fastest, most performant speech recognition and voice synthesis models.

From customer support to drive-thru agents, developers can quickly and easily build AI voice agents that listen, think, and respond in real-time, as naturally as a human can.

The API provides:

✔ Natural-sounding conversations

✔ Smooth interruption handling

✔ More control – choose from open source, closed source, or BYO LLM

✔ Low cost to scale with confidence

Sign up to request access.

Tesla unveils Optimus upgrades

When Elon Musk first unveiled his humanoid Optimus robot in 2021, Tesla had yet to make much progress. At the time, all they had was a man in a suit. The next year, they had hardware, but it didn’t really do anything. 

But since 2023, progress has been (seeming) to move at a far quicker pace. 

The details: Tesla on Thursday unveiled a series of upgrades made to its latest generation of Optimus robots in a brief video posted to Twitter. 

• The robot, according to Tesla, can now explore “unseen” spaces autonomously, navigating obstacles through the neural networks that power it. 

• It can also climb stairs and autonomously carry around heavy payloads without overheating. 

“Our work on Autopilot has greatly boosted these efforts; the same technology is used in both car & bot, barring some details and of course the dataset needed to train the bot’s AI,” an Optimus engineer said. 

Of course, these produced demos aren’t a clear indicator of progress; we don’t know how many takes it took to get the robot to walk around properly, and we don’t know if/how remote human intervention was used in the production of the video. What we do know is that plenty of remote intervention was used at Tesla’s recent robotaxi event, giving the misleading appearance that the Optimus bots could make drinks and chat off-the-cuff with attendants. 

Learn from Nvidia, HP, Databricks, and more

Learn actionable insights, strategies, and techniques for generative AI stack design, governance, evaluation, and observability.

Register now for GenAI Productionize 2.0 – the premier conference for GenAI application development, featuring AI experts from leading brands, startups, and research labs!

But don’t take our word for it; here are real quotes from previous attendees:

"I'm blown away by the high quality and value of this event." - Ricardo B.

"Great event - worth getting up at 4am in the morning for!" - Sandy A.

"Spectacular and very insightful summit! Very well done!" - Chad B.

• Retail startup Autone raised $17 million in Series A funding.

• Search startup Tako raised $5.75 million in seed funding.

• 78 AI startups that could be up for sale this year (The Information).

• Biden administration forgives $4.5 billion in student debt for over 60,000 public service workers (CNBC).

• EU backs nuclear power ahead of COP29 (Semafor).

• Anthropic’s CEO thinks AI will lead to a utopia — he just needs a few billion dollars first (The Verge).

• Bain & Co, OpenAI expand partnership to sell AI tools to clients (Reuters).

If you want to get in front of an audience of 200,000+ developers, business leaders and tech enthusiasts, get in touch with us here.

Hiring the best Machine Learning Engineer just got 70% cheaper.

• Today we are highlighting AI talent for you, courtesy of our partner, Athyna. If you are looking for the best bespoke tech talent, these stars are ready to work with you—today! 

• Reach out here if we can make an introduction to these talents and get a $1,000 discount because you’re a reader of The Deep View!

• Trag: A tool to auto-review code and suggest fixes based on custom natural language rules.

• Guidde: A tool to create video documentation.

TSMC spikes on ‘monster’ third quarter

Taiwan Semiconductor Manufacturing Company (TSMC) — the firm that supplies companies like Nvidia with advanced AI chips — reported strong third-quarter earnings results on Thursday which kicked off a significant rally for its U.S.-traded stock. 

Shares of TSMC closed Thursday up around 10%. Nvidia jumped around 2% in kind.

The details: TSMC reported a 36% year-over-year increase in revenue and a 54% increase in net income for the quarter. 

• The semiconductor brought in $23.5 billion in revenue for the quarter, on “strong smartphone and AI-related demand.”

• The firm expects that demand to continue into the fourth quarter of the year; TSMC guided revenue of at least $26.1 billion for the current quarter. 

The semiconductor side of things is where we’ll be able to track the health of the AI sector and the AI trade. So long as chip demand remains high and growing, things are looking good. By that same token, if the chipmakers start reporting flagging or decreasing demand, that could be a sign that the industry is in distress. 

During Thursday’s earnings call, TSMC Chairman and CEO C.C. Wei said that AI demand is “real,” adding: “Almost every AI innovator is working with TSMC.”

Wedbush analyst Dan Ives said that TSMC’s “monster” numbers are a “very important data point for the AI Revolution thesis and growth looking ahead … TSMC is what matters for … the AI trade.”  

Competition and power consolidation in AI

Since the ‘AI race’ kicked off at the beginning of 2023, there has been a concern — echoed in policy debates — about a suffocating consolidation of power in the industry. It’s not really a surprise; despite the billions of venture capital funding flowing into AI startups, it is enormously expensive to train and operate AI technology, something that has positioned Big Tech actors to easily dominate in the sector. 

Firms like Google, Amazon, Meta and Microsoft — companies that each earn hundreds of billions in revenue each year and are collectively valued at around $8.5 trillion — have the cash at their disposal, from their existing businesses, to fund expensive ventures in artificial intelligence. 

But for many startups, and importantly, for universities, the creation and study of in-house models is largely cost-prohibitive, something that has already entrenched Big Tech dominance.

• One indication of this cost disparity — and the control exercised by Big Tech — involves the numerous “aqui-hires” that some of these firms have completed. 

• Microsoft, for example, hired the founders and most of the staff of Inflection AI this year. Google made a similar deal with Character. And in July, Amazon made a similar deal with Adept. 

An addition to this environment is the lack of competition around the necessary hardware; TSMC and Nvidia are at the center of it all. Their competitors are few and far smaller. 

A recent report from Mozilla and the Open Markets Institute — ‘Stopping Big Tech from Becoming Big AI’ — explores the ecosystem, making a number of policy proposals that would help level the playing field. 

“While most of these practices are targeted at competitors rather than end users, by reducing competition in AI, they risk undermining innovation and leaving consumers with fewer options – including fewer options that put their privacy and safety first,” the report reads. 

The remedies: Arguing that governments already have the legal authority to rein the sector in, the report’s most significant recommendations center around the regulation of the dominant players. 

• This includes anti-trust monitoring and enforcement, as well as regulation to ensure that the market remains diverse and that the necessary inputs remain open and accessible. 

• The report also recommends far more stringent merger control to stymie the kinds of acquisitions and investments that Big Tech firms deploy as a kind of strategy to waylay competition. 

The report additionally recommends that governments pursue regulation that would require transparency — “it is not enough to understand how an AI system works; there must also be equitable access to these systems.” This, according to the report, would ensure that AI products are safe while also preventing monopolies from growing up in the space. 

“This potentially globally disruptive technology is too important to be left wholly at the whims of today’s tech monopolists,” the report says.

We often talk about the risks of AI in the context of control. The fears around general or super intelligence are based on concerns of controllability. 

My concerns also center around control. But they have nothing to do with general intelligence and everything to do with the economic context in which these current systems exist. 

Right now, we have no control over what Big Tech chooses to do with AI. Whether that means supplying militaries with fundamentally flawed systems, or inserting unwanted, unsustainable generative software into platforms that worked just fine before, or scraping all of our data to build (and then sell us) generative AI models, Big Tech does what it wants. 

And in many ways, the sector is already entrenched through years of dominance in the internet and social media.

Regulation is the only path toward establishing some element of control. Control is the only way we can have some semblance of hope that AI will be deployed equitably, in ways that at the bare minimum, don’t cause harm. 

🤔 Your thought process:

Selected Image 2 (Left):

• “The coffee beans weren’t that good on the fake photo.”

Selected Image 1 (Right):

• “It looked like an influencer photo on Instagram.”

💭 A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Here’s your view on AI & nuclear power:

Around 40% of you think it’s fantastic; 20% are worried because it’s Big Tech and 16% said it can’t be worse than fossil fuels.

The rest aren’t sure.

I wish I had space to include more of your thoughts on this.

Worried:

• “I support the implementation of nuclear energy, and it is great that Big Tech is investing, but it is because they are straining the grid with their AI products, and I don't trust their greed around the necessary safety of nuclear energy.”

Something else:

• “While resurgence of nuclear power use may be inevitable, the safety concerns and waist storage problems need to be aggressively addressed across the board.”