Bleak and Bright

The Angle Issue #225

Bleak and bright
David Peterson

We’re at an interesting moment in technology history. Artificial intelligence is here, getting more powerful by the second. But the future remains, as always, maddeningly opaque. I imagine this may be what it felt like during the earliest days of the internet, or at the beginning of the Atomic age. Depending on the day, it can feel equally likely that a world with increasingly powerful artificial intelligence will be suffocatingly bleak or blindingly bright.

Here’s an example of what I mean.

Play around with Suno or Sora or any of the LLM models and it’s not hard to imagine that we soon enter a world of near-infinite, hyper-personalized, algorithmically-generated, content. Music, books, videos and games, custom developed for you, and beamed directly to your phone or your Meta Ray-Bans or your Apple Vision Pro. Everything and anything generated on-the-fly with one purpose: to keep you plugged in.

How will “real life,” which is beautiful but often quite frustrating and boring, compete with an infinite scroll myopically designed to keep us all coming back for more? Perhaps we’re closer to a Matrix-style plugged in future than any of us would like to think.

That’s one argument. But here’s a counter. What if we build, as a friend of mine calls them, personal AI-powered dopamine defense systems? For example, what if I deploy a personal AI agent that consumes digital content for me and highlights only the content that promotes my personal cognitive development and deepens my learning related to specific topics? (Kind of like that personal AI that goes on dates for you to find the ideal match).

That actually sounds pretty amazing, right?

This is a social/consumer internet example, but the exact same dynamic is playing out in B2B.

We’re seeing a deluge of AI-powered sales tech right now. AI-powered lead generation. Automated lead qualification. Hyper-personalized outreach, written on the fly without a salesperson in sight. Buyers are getting inundated across the board. But, from what I’ve heard, the results aren’t as bad as you might think. We aren’t as good at telling human from AI as we’d like to believe…at least not yet.

So what’s the counter? Well, I’m sure soon enough we’ll have AI agents for our work email inboxes just like for our social feeds. We’re in the midst of an AI arms race, and our inboxes may be the first battleground on which these agents fight.

In reaction, one portfolio company has decided to lean hard into the partner-powered co-sell motion. They’re building a platform that orchestrates the incredibly complex process of getting multiple stakeholders from across companies aligned for a single sale (see here). Co-selling is challenging, but their bet is that, as inboxes become wastelands, buyers will increasingly want to buy from people they know and trust. It’s early, but I’ve started to see this trend play out in some of the more sophisticated GTM organizations already.

It’s clear that artificial intelligence will be used to boost the noise in all of our lives at the expense of signal. And, for what it’s worth, I don’t doubt that at least some percentage of the population will fall prey to the siren song of the hyper-personalized infinite scroll. But I remain firmly on the side of the personal AI rebels. And if that describes what you’re building…let’s talk.



Revenue Durability in the LLM World
Everything about LLMs seems to make revenue durability more challenging than ever.

A Digital Fabric for Maritime Trade
Why we invested in Portchain.

Three Keys to the Kingdom
The sometimes-competing and sometimes-aligned goals that early-stage founders must manage.


UK / Artificial Intelligence. Wayve raised $1B, led by Softbank, to continue development of its “embodied AI” autonomous driving platform for Tesla competitors.

UK / FinTech. Abound raised up to £800M (in debt and equity), led by GSR Ventures, to scale its personalized loan platform.

UK / FinTech. Monzo raised $190M, led by Hedosophia and CapitalG, to fund US expansion of the UK-based neobank.

Israel / DefenseTech. XTEND raised $40M, led by Chartered Group, for its drone training and operations platform.

UK / FinTech. Fintech Farm raised $32M, led by Nordstar, to expand its “neobank-in-a-box” platform to India.

Denmark / Manufacturing. NIL Technology raised $31M, led by Export & Investment Fund of Denmark, Jolt Capital, NGP Capital, and Swisscanto Private Equity, to scale its nanoimprint lithography manufacturing capabilities.

Israel / Security. LayerX raised $26M, led by Gilot Capital Partners, for its browser-based security platform.

Israel / SaaS. Fairgen raised $8M, led by Maverick Ventures Israel, for its AI-powered synthetic audience survey platform.



What’s next in AI chips. The MIT Technology Review covered the ongoing battles in the chip wars. One major battleground is edge computing, which will enable better data privacy for AI applications. “Currently, most of our interactions with AI models like ChatGPT are done via the cloud. That means that when you ask GPT to pick out an outfit (or to be your boyfriend), your request pings OpenAI’s servers, prompting the model housed there to process it and draw conclusions (known as “inference”) before a response is sent back to you. Relying on the cloud has some drawbacks: it requires internet access, for one, and it also means some of your data is shared with the model maker. That’s why there’s been a lot of interest and investment in edge computing for AI, where the process of pinging the AI model happens directly on your device, like a laptop or smartphone. With the industry increasingly working toward a future in which AI models know a lot about us (Sam Altman described his killer AI app to me as one that knows “absolutely everything about my whole life, every email, every conversation I’ve ever had”), there’s a demand for faster “edge” chips that can run models without sharing private data. These chips face different constraints from the ones in data centers: they typically have to be smaller, cheaper, and more energy efficient.” But here, as elsewhere in semiconductor land, it’s incredibly difficult for small startups - however well funded - to displace the majors. The piece interviews Murat Onen, an entrepreneur who is working to develop a chip based on a new architecture. “One day in the lab, “through optimizing these numbers, and getting very lucky, we got the material that we wanted,” Onen says. “All of a sudden, the device is not a science fair project.” That raised the possibility of using such a component at scale. After months of working to confirm that the data was correct, he founded Eva, and the work was published in Science. But in a sector where so many founders have promised—and failed—to topple the dominance of the leading chipmakers, Onen frankly admits that it will be years before he’ll know if the design works as intended and if manufacturers will agree to produce it. Leading a company through that uncertainty, he says, requires flexibility and an appetite for skepticism from others.”

ARM is also jumping in AI. According to this report in The Information, UK-based chipmaker ARM is also planning to join the AI semiconductor fray. “According to the report, U.K.-based Arm would be responsible for initial development costs related to the chip, but SoftBank Group, which owns a majority stake in Arm, could also contribute. If the AI chip becomes ready for mass production, it could be spun off under a new SoftBank subsidiary. Bloomberg reported earlier this year that Masayoshi Son, the founder of SoftBank, was seeking up to $100 billion to fund an AI chip company. SoftBank is already talking to Taiwan Semiconductor Manufacturing Company and others about securing manufacturing capacity, according to the Nikkei report. SoftBank has been looking for ways to help Arm benefit from the boom in AI computing. The company designs chips known as central processing units, which are needed for AI computing but far less so than graphics processing units. Arm’s limited AI exposure has held the company back as rival Nvidia, the predominant maker of GPUs, records blockbuster revenue growth.”

HuggingFace moves into robotics. France’s Hugging Face, which rocketed to fame and success as the “github for machine learning” is moving forcefully into the robotics world. The insight that AI will be a critical unlock for robotics is increasingly widespread: AI can keep robotics costs down by enabling standard robotics physical interfaces to adapt to a wide variety of real-world workloads. According to ZDNet, “Hugging Face appears to be the latest AI player joining the robotics race, as a former Tesla scientist says he's starting an "ambitious" open-source robotics project at the AI tech platform. "After 3 years @tesla and Optimus, I am thrilled to announce that I joined Hugging Face to start an ambitious open robotics project! (open as in open-source, not as in Open AI) Looking for engineers to build real robots in Paris," AI researcher Remi Cadene announced on X, formerly Twitter.”

The trouble with rare earth magnets. The Wall Street Journal dug deep into how the US defense industry (among other industries) depends on rare earth magnets that are largely sourced in China - and how the US government is trying to incentivize the private sector to close the gap. “The Defense Department in the past few years has committed more than $450 million toward rare earths and the magnets they power. The Energy Department is offering its own incentives because the magnets are also critical for electric vehicles. The funding is helping a German magnet-maker set up its first North American factory, which broke ground in March, two decades after its last U.S. factory shut down. The facility, in Sumter, S.C., will buy rare earths locally. Those supplies could come from other projects that are receiving government funding—such as processing plants coming up in California and Texas, owned by American and Australian miners, respectively. Their highest hurdle is low Chinese prices. A U.S. Commerce Department probe in 2022 found that China’s dominant position enabled it to set prices low enough to make production unsustainable for competitors.”


Don’t be afraid of unpopular VPs of Sales. Jason Lemkin of SaaStr make a compelling case that a well-liked candidate for VP of Sales may not be the best one. It’s worth a read. “Your entire existing sales team should NOT love a great VP of Sales candidate. Why? First, the lower performers will be threatened by someone great. They know they’ll likely be moved out. And also, top performers may be threatened as well. Why? Because things will likely at least change. And top performers are usually comfortable with the way things are, at least for them. The best VPs of Sales immediately identify the top performers and make them want to stay when they join. But this doesn’t eliminate all anxiety or change. I almost always see the top performing IC sales execs at a startup preferring a “nice” VP of Sales that often doesn’t know the product, pace or industry — but will support them and leave them be. Logical, but usually the wrong choice for the company. Still, if they are 100% opposed to a candidate, that is a sign to generally not make the hire.”

Public speaking tips. Polina Pompliano compiled a list of public speaking tactics that have helped her. This can be a challenge for many founders, and her tips are helpful. “If you struggle with nerves, you need to replicate that feeling of having your heart beat out of your chest while delivering your speech. Artist Taylor Swift recently said that she trained for her “Eras” world tour by singing her entire setlist while running on the treadmill. She did this every day for six months before the tour kicked off. Why? She wanted to mimic the feeling she would have on stage and make sure she never ran out of breath while performing for four straight hours. The same idea can be applied for any public speaking situation. Your body will respond on its own — especially if you’re someone prone to nerves. When I’m nervous, my heart starts racing, I start sweating, and my voice starts shaking. So how can I artificially summon these physical symptoms and practice my speech under duress? I’ve found that doing a moderate to intense workout (like a run on the treadmill) a few hours before my speech really makes a difference. As I start my run, my heart rate spikes and I start sweating. At that point, I start practicing my speech to make it muscle memory, allowing my brain to know that I can calmly get through it no matter how much physical stress I’m under.”


Raising a new VC fund continues to be harder than ever. Techcrunch spoke with a number of emerging GPs and LPs, including Joanna Drake of Magnify Ventures. “Most of the action is taking place between zero and $49 million, where roughly 50% of emerging managers are raising, Drake said. “That’s important because while there’s a handful of emerging managers that are able to raise larger than $100 million funds, it’s really a small percentage of the market,” Drake said. “So, they actually do not have the capital to take the companies to a later stage. They have to work with the larger firms and put together the syndicates. It’s actually one of the most important roles that they play.” And, even if emerging fund managers successfully deploy their first funds and have good early results to show (although most funds take 10 years to return), that’s not enough to be secure. Theresa Hajer, head of U.S. venture capital research at Cambridge Associates, agrees that there’s been an influx of emerging manager funds over the past seven years…Cambridge is to VC funds what Michelin is to restaurants, helping to identify the best performers. But because of the odd winter period we’re in, past success isn’t actually a strong indicator on its own to access emerging managers, she warns. Newer managers who were investing during the 2019-2021 party days haven’t yet had the opportunity to build a track record in an environment that has had a valuation reset. So limited partners “need to sharpen their pencils and look very carefully because you can’t always rely on that performance,” she said.”

On keeping venture small. Nico Wittenborn, founder of Adjacent Ventures, gave a very thoughtful interview to Patrick O’Shaughnessy on Invest Like The Best. He talks about consumer subscription businesses and some of his investment theses, but the best part of the conversation was his observations on the venture capital market - specifically on the benefit of keeping venture firm teams small - sometimes very small. He emphasizes the importance of partners doing their own due diligence, reaching their own investment convictions, not outsourcing their decision-making to others, and building direct deep relationships with founders.


Reco was named a winner at the 12th Annual Global Infosec Awards 2024.

JFrog announced an all-star speaker line-up for swampUP 2024.



or to participate.