Germanys Aleph Alpha has raised over $500 million in a Series B round backed by Bosch and SAP among other new investors, as the startup looks to build on its promise to be the leading provider of sovereign generative AI applications in Europe, taking on OpenAI and the hyperscalers.

Aleph Alpha, founded by former Apple AI researcher Jonas Andrulis, has created the Luminous series of large language models (it has promised a 300 billion parameter Luminous World model later this year) which are commercially available for customers now, via a Python client.

Working with partner Graphcore, it has demonstrated some impressive AI compute efficiency progress as well as capabilities with a privacy and explainability focus (see for example its Atman paper of earlier in 2023.)

The Series B funding, it said in a press release on Monday, strengthens the foundation for Aleph Alpha to further advance its proprietary AI research, accelerate development and commercialization of Generative AI for the most complex and critical applications such as in data sensitive industries like healthcare, finance, law, government and security.

At a press conference on Monday, Germanys Minister for Economic Affairs Robert Habeck suggested that the investment played to Europes strategic national priority as work continues to boost its data sovereignty.

The thought of having our own sovereignty in the AI sector is extremely important. If Europe has the best regulation but no European companies, we havent won much Habeck said at the press conference.

Aleph Alpha will continue to expand its offerings while maintaining independence and flexibility for customers in infrastructure, cloud compatibility, on-premise support and hybrid setups said CEO Andrulis.

The ongoing developments will extend interfaces and customization options tailored to business-critical requirements he added in a release.

The massive funding round came as researchers at Google DeepMind suggested in a widely shared paper that the transformer models powering so much of the past years AI hype were not as intelligent, perhaps, as many seem to believe: When presented with tasks or functions which are out-of-domain of their pretraining data, we demonstrate various failure modes of transformers and degradation of their generalization for even simple extrapolation tasks they wrote in a paper on November 3.

Together our results highlight that the impressive ICL [in-context learning] abilities of high-capacity sequence models may be more closely tied to the coverage of their pretraining data mixtures than inductive biases that create fundamental generalization capabilities.

Link:
EU investors including Bosch, SAP pump $500 million into ... - The Stack

Last week was an important moment in the debate about AI, with President Biden issuing an executive order and the UKs AI Safety Summit convening world leaders.

Much of the buzz around these events made it sound like AI presents us with a binary choice: unbridled optimism, or existential fear. But there was also a third path available, a nuanced, practical perspective that examines the real risks and benefits of AI.

There have been people promoting this third perspective for years although GPT-fueled headlines of the past 12 months have often looked past them. They are foundations, think tanks, researchers and activists (including a number of Mozilla fellows and founders) plus the policymakers behind efforts like last years AI Blueprint for an AII Bill of Rights.

We were happy to see the executive order echo many of the ideas that have emerged from this school of thought over the last few years, prioritizing practical, responsible AI governance. The UK Safety Summit started on a very different note, anchored in concerns around existential risks but also some welcome reframing.

As we look forward from this point, it feels important to highlight three key levers that will help us get closer to responsible AI governance: well-designed regulation, open markets, and open source. Some of these were in the news last week, while others require more attention. Together, they have the potential to help us shape AI in ways that are more trustworthy, empowering and equitable.

Regulation

As we saw last week, there is near consensus that AI presents risks and harms, from the immediate (from discrimination to disinformation) to the longer term (which are still emerging and being explored). Theres also a growing consensus that regulation is a part of the solution.

But what exactly this regulation looks like and what its outcomes should be is where consensus breaks down. One thing is clear, though: Any regulatory framework should protect people from harm and provide mechanisms to hold companies accountable where they cause it.

The executive order included encouraging elements, balancing the need for a rights-respecting approach to addressing AIs present risks with exploration of longer-term, more speculative risks. It also acknowledges that the U.S. is still missing critical baseline protections, such as comprehensive privacy legislation that work hand-in-hand with AI-specific rules.

The ideas that dominated the Safety Summit were less encouraging. They reinforced that old binary, either going too far or not far enough. There was a focus on self regulation by AI companies (which isnt really governance at all). And there were nods towards the idea of licensing large language models (which would only increase concentration and may worsen AI risks, in the words of Sayash Kapoor and Arvind Narayanan).

Open markets

To Arvind and Sayashs point, there is a problematic concentration of power in the tech industry. Decisions about AI, like who it most benefits or who is even allowed to access it, are made by a handful of people in just a few corners of the world. The majority of people impacted by this technology dont get to shape it in any meaningful way.

Competition is an antidote. AI development not just by big companies, but also smaller ones (and nonprofits, too) has the potential to decentralize power. And government action to hinder monopolies and anti-competitive practices can accelerate this. The executive order takes note, calling on the Federal Trade Commission (FTC) to promote competition and protect small businesses and entrepreneurs.

Its important for this work to start now both by enforcing existing competition law, but also by greater adoption of ex-ante interventions like the UKs DMCC bill. The previous decade showed how quickly incumbent players, like social media platforms, acquire or shut down competitors. And its already happening again: Anthropic and OpenAI have familiar investors (Google + Amazon and Microsoft, respectively), and once-independent laboratories like DeepMind were long ago acquired (by Google).

Open source

For smaller AI players to thrive in the marketplace, the core building blocks of the technology need to be broadly accessible. This has been a key lever in the past open-source technology allowed a diverse set of companies, like Linux and Firefox, to compete and thrive in the early days of the web.

Open source has a chance to play a role in fueling competition in AI and, more specifically, large language models. This is something organizations like Ai2, EleutherAI, Mistral, and Mozilla.ai are focused on. Open source AI also has the potential to strengthen AI oversight, allowing governments and public interest groups to scrutinize the technology and call out bias, security flaws, and other issues. Weve already seen open source catch critical bugs in tooling used for core AI development. While open source isnt a panacea and it can be twisted to further consolidate power if its not done right it has huge potential in helping more people participate in and shape the next era of AI.

Its important to note that there is a major threat to open source AI emerging: some use the fear of existential risk to propose approaches that would shut down open-source AI. Yes, bad actors could abuse open source AI models but internet history shows that proprietary technologies are just as likely to be abused. Rushing to shut down open source AI in response to speculative fears, rather than exploring new approaches focused on responsible release, could unnecessarily foreclose our ability to tap into the potential of these technologies.

Collaboratively dealing with global problems is not a new idea in technology. In fact there are many lessons to learn from previous efforts from how we dealt with cybersecurity issues like encryption, governed the internet across borders, and worked to counter content moderation challenges like disinformation. What we need to do is take the time to develop a nuanced approach to open source and AI. We are happy to see the EUs upcoming AI Act exploring these questions, and the recent U.S. executive order instructing the Department of Commerce to collect input on both the risks and benefits of dual-use foundation models with widely accessible weights in essence, open-source foundation models. This creates a process to develop the kind of nuanced, well-informed approaches we need.

Which was exactly the goal of the letter on open source and AI safety that we both signed last week along with over 1,500 others. It was a public acknowledgement that open source and open science are neither a silver bullet nor a danger. They are tools that can be used to better understand risks, bolster accountability, and fuel competition. It also acknowledged that positioning tight and proprietary control of foundational AI models as the only path to safety is naive, and maybe even dangerous.

The letter was just that a letter. But we hope its part of something bigger. Many of us have been calling for AI governance that balances real risks and benefits for years. The signers of the letter include a good collection of these voices and many new ones, often coming from surprising places. The community of people ready to roll up their sleeves to tackle the thorny problems of AI governance (even alongside people they usually disagree with) is growing. This is exactly what we need at this juncture. There is much work ahead.

Excerpt from:
A Third Way on AI - Mozilla & Firefox

Dive Brief:

Google was an early incumbent to enter the race to embed generative AI into core tools and services. As more providers launched their own capabilities, simply offering generative AI stopped being innovative. Its expected.

Its almost irresponsible for a vendor not to have AI in their literature somewhere or theyre irrelevant all of the sudden, said Greg Myers, operating partner at investment firm Cota Capital, speaking Tuesday during the Dell Technologies Forum in Washington, D.C.

Enterprises are in search of the best of the best when it comes to tooling, but are taking different approaches to get there. Some are leaning on their existing cloud providers, drawn by the comfort of familiarity. Others are deploying crowdsourcing methods to empower employees, while another portion is waiting for the cream to rise to the top.

Google expects its array of solutions to win over customers as competition mounts. Currently, more than half of all the funded generative AI startups are Google Cloud customers, including AI21 Labs, Contextual, Elemental Cognition and Rytr, according to Pichai.

I view it as a journey, and each generation is going to be better than the other, Pichai said. We are definitely investing and the early results are very promising.

The companys CapEx costs increased in Q3, reaching $8 billion from $6.9 billion in the previous quarter. The growth was driven overwhelmingly by Googles technical infrastructure enhancements to support compute-heavy AI workloads, CFO Ruth Porat said. Upgrading infrastructure in preparation for increased adoption of generative AI is a trend across the cloud hyperscalers, including AWS, Microsoft Azure and Oracle.

We remain committed to durably re-engineering our cost base in order to help create capacity for these investments in support of long-term sustainable financial value, Pichai said. Across Alphabet, teams are looking at ways to operate as effectively as possible, focused on their biggest priorities.

See more here:
Google lays foundation of next-stage AI with Gemini LLM - CIO Dive

And you thought Devilsdorp was disturbing Stewart Boetie Boer Wilken is a South African serial killer who was active in Port Elizabeth (now Gqeberha) during the 1990s. In February 1998, he was convicted on seven counts of murder (and two of sodomy, a few weeks before gay sex was legalised in South Africa) but admitted to numerous more murders in his written confession.

This Showmax Original true-crime series is the most frightening to date.Director Jasyn Howes interviews the detectives and forensic psychoanalysts involved in Wilkins capture, one of the victims family, and most alarmingly, the series takes us back to 1990 recreating Wilkins killing spree based on a recorded interview with the murderer himself from his prison cell.

The interview took place in 2006 with Dr Grard Labuschagne, the former section head of the Investigative Psychology Section of the South African Police Services, who also features in the series.

Labuschagne gave us his recording of just over three hours of Wilken sharing, in his own words, everything hed done, says Howes.

Wilken had been in prison for eight years at the time of that interview, and was already sentenced to life behind bars, so he had no obvious reason to hide anything.

It should go without saying that a documentary with re-enacted crimes and first-person accounts of a ruthless serial killer will disturb some people, but best to say it anyway: things get evil very quickly this is not an easy watch.

Dont think because you make it through the first episode without nightmares that youll handle the rest. It only gets more intense as Wilkins reveals how his fiendish crimes escalated. Murder, rape paedophilia, necrophilia, bestiality, suicide, cannibalism, incest you name it, this story has it all.

Even if you are over 18 and a fan of true-crime documentaries, we strongly advise viewer discretion, says Allan Sperling, executive head of content at Showmax.

There have been a lot of conversations recently about the dangers of glamorising true crime in pop culture. When production houses leverage audiences fascination (some would say morbid fascination) with murder to profit off of real-life horrors, this can re-traumatise victims and potentially even inspire copycat criminals.

In 2022, Netflix came under fire for creating a series called Monster: The Jeffrey Dahmer Story, fictionalising the crimes of a real serial killer without consulting the families of the victims he brutally murdered. Even crime documentaries may hinder the pursuit of justice in some cases. Howes has made a concerted effort to avoid sensationalism and ethical pitfalls. Each episode is bookended by disclaimers indicating his awareness of these debates and stating his position on them. Episodes start with the text:

The documentary aims to shed light on the chilling complexity of human actions and the pursuit of justice and not to glorify crime, and end with Reasonable efforts were made to contact the families of the victims of Stewart Wilkens crimes.

The subtitle of this series speaks directly to the reason true crime is so popular many people are extremely curious about how someone can do such terrible things, and what made them that way. These were the questions clinical psychologist Gerard Labuschagne sought to answer when he interviewed Wilken in 2006, to aid in profiling serial killers in general.

This pursuit is plagued by confirmation bias. Because the prospect of understanding criminals better is one of the strong appeals of true crime, serial killer documentaries tend to fixate on causality. If Wilken says it was his being bullied as a child that caused him to become a murderer, the filmmakers seem keen to believe him because it benefits their storytelling. But, as Labuschagne hints in the series, its possible that this explanation is just an attempt to justify his unjustifiable actions by attributing them to his sob story.

Clinical psychologist Giada Del Fabbro builds on this theory in the series,quoting psychological theorist, William Fairburn who said A child would rather be the devil in a world where god existed than have no god whatsoever. Its a compelling notion, and yet loads of people are bullied or abused and still dont grow up to become serial killers, so its problematic to latch onto this explanation.

The truth is that we can never truly get inside the mind of a monster. That is not to say that we shouldnt try, for the purpose of spotting people from becoming criminals, but we should be careful about jumping to conclusions about how minds work just because it aligns with the way we understand the world.

A snippet of Labuschagnes interview in the series distils what makes serial killers such a frightening idea its not that theyre so different from the rest of us, its actually that theyre so similar:

When you meet these people, you only know them through the context of the crimes theyve committed, so whenever you do meet them its not what you expect. 99.9 per cent of the time theyre doing the same things everyone else is doing.

Serial killers often take advantage of positions of trust to mask their crimes. Wilken didnt exactly have the charisma of Ted Bundy, but he was purportedly great with kids and lived in a coloured community, spoke the language and the slang, and had married a woman of colour in a time when that was very unusual for a white Afrikaans man. The mere fact that he was known by the term of endearment Boetie is an indication that there was trust in the community which he could exploit.

The 1990s in South Africa was a time of euphoria. The country was focused on Nelson Mandelas release, its first free elections, the 1995 Rugby World Cup etc. Crime reporter Brett Adkins notes in the series that these fantastic changes somewhat distracted the media from layers of evil that remained.

During apartheid, the police force had been focused on other kinds of crimes, and law enforcement was not trained in serial killing. This meant serial murder could easily be missed because it wasnt a priority. Norman Simons, the infamous Station Strangler, was only convicted in one case and was released on parole in July 2023.

It was only the urgency of SAPSs inability to detain Moses Sithole, The ABC Killer, while he was actively murdering that led them to enlist the help of an ex-FBI forensic psychologist to train them to create serial killer profiles. This training was essential in catching Wilken a few years later because he defied some of the usual profile trends.

Wilken was a highly unusual serial killer, says Howes, Unlike most serial killers, he had more than one type of victim: predominantly female sex workers and young boys, usually street children, across multiple races.

Despite committing at least seven murders, Wilken didnt receive as much coverage as youd expect, and this is generally attributed to the countrys focus on its large political changes.

Its also interesting to discover how the context of South Africa in the 90s shaped Wilkens descent into criminality personally. Wilken was a fisherman in a period when the industry was highly intertwined with drug trafficking. Drug boats would sell stimulants to fishermen to keep them working through the long hours of a haul employees would work on trips of a week or two straight and then be paid in a lump sum, meaning that low and middle-income communities would often have drug users return home loaded with cash to feed their habit. This vicious cycle exacerbated Wilkens psychopathy and aggression and spurred his first murder.

Because there was almost no video or radio archive to draw on, the series leans into recreations to bring the story to life visually. Theyre not great, and they get progressively worse. To put distance between the present and these past events, the re-enactments employ an overly theatrical technique showing the clips in slow motion, muting them, and overlaying dialogue with a reverb effect.

Coupled with a yawn-inducing eerie soundtrack, most of the re-enactment scenes have the dull vibe of a try-hard ghost story, with the exception of Raven Swarts scenes as a young Wilken, which have an appropriately spine-chilling effect.

The interview sections of the film are less sensational and more gripping. The series explores the typical themes that serial killer shows go for, like early development and the nature of psychopathy, and having all that discussed by familiar types of South African characters brings it closer to home. DM

Boetie Boer is available on Showmax. Episodes air weekly on Wednesdays.

You can contact What Were Watching via[emailprotected]

Follow this link:
'Boetie Boer': a deep dive into the terrifying mind of a monster - Daily Maverick

The companys technology, which removes the engineering complexity around AI systems for image analysis, will initially assist medical professionals

Founder co-created Madbits, a deep-learning image-analysis start-up that was acquired by Twitter

PARIS & NEW YORK, October 26, 2023--(BUSINESS WIRE)--Lcole AI, creator of O, a "machine teaching" technology that removes the engineering complexity around deep-learning systems for computer vision, announced that it has raised 3 million USD in Seed funding from Sofinnova Partners. The funds will be used to develop the companys proprietary technology, which will be rolled out to medical professionals and researchers, enabling them to create bespoke AI systems to assist in their area of expertise. In addition, the financing will fuel the expansion of the team, accelerate product development and structure business efforts. Also participating in the round are notable business angels including Preston-Werner Ventures, the fund started by co-founder and former CEO of GitHub, Tom Preston-Werner.

"We're building a user interface for AI so anyone can create and benefit from their own personalized AI assistant," said Louis-Alexandre Etezad-Heydari, Co-Founder and President of Lcole AI, which means "AI school" in French. "We also are creating a system that will open possibilities for secure collaboration between organizations."

Etezad-Heydari co-founded Madbits, a deep-learning image-analysis start-up, with Clment Farabet in 2013. A year after its founding, Madbits was acquired by Twitter, where the two entrepreneurs ran Twitter Cortex, an internal team that built a Deep Learning platform to power recommendation systems, search, ranking and filtering at Twitter.

Kim Nilsson and Jonathan Alexander Brown teamed up with Etezad-Heydari to perfect the AI development framework for computer vision that initially inspired Madbits, and the three founders set off with the goal to make creating computer vision models simple enough for non-engineers.

Story continues

"L'coles technology is designed to democratize machine learning by enabling life science researchers and other non-machine learning experts to utilize tailored computer vision systems, thus accelerating life sciences research," said Edward Kliphuis, Partner at Sofinnova Partners. "The focus on digital medicine is a logical entry point," he noted.

"Were starting with a focus on health care and life sciences because we want to make a positive impact right away," said Jonathan Alexander Brown, Co-Founder and Chief Executive Officer. "With Sofinnova's support, we are confident we have the right skills to partner with researchers and clinicians in these tightly regulated markets."

Farabet, an investor in Lcole, is an AI pioneer. Currently VP of Research at Google DeepMind, he spent six years as a senior executive at NVIDIA, working on its autonomous vehicles and the company's data science platform. Farabet is also famous in the AI world as one of the creators of Torch, a machine learning framework that provides a simple and flexible interface for building and training deep neural networks.

Lcole AI counts a number of other AI pioneers among its investors, including Nicolas Pinto, head of Deep Learning at Apple, Clment Delangue, Co-Founder and CEO at Hugging Face.

About the Founders

Jonathan Alexander Brown, Co-Founder and CEOA mathematician-turned-actuary, Jonathan has crafted statistical models for natural catastrophes (wildfires, hurricanes, earthquakes, etc.) and introduced innovative risk management and data-analysis strategies at top multinational companies. He has navigated regulated environments throughout his career, always with a keen focus on addressing the needs of key stakeholders.

Louis-Alexandre Etezad-Heydari, Co-Founder and PresidentLouis-Alexandre is a computer vision pioneer, a successful entrepreneur, and an accomplished artist who dropped out of the Neuroscience PhD program at New York University to found Madbits in 2013 with Clment Farabet. The pair "built visual intelligence technology that automatically understands, organizes and extracts relevant information from raw media," Etezad-Heydari said when Madbits was acquired by Twitter.

Kim Nilsson, Co-Founder and CTOKim has more than 20 years' experience as a software engineer. A former senior developer at Opera, he has always displayed the hacker spirit, solving one of the biggest cryptocurrency heists ever in his spare time, an effort the Wall Street Journal described in detail.

About Sofinnova PartnersSofinnova Partners is a leading European venture capital firm in life sciences, specializing in healthcare and sustainability. Based in Paris, London and Milan, the firm brings together a team of professionals from all over the world with strong scientific, medical, and business expertise. Sofinnova Partners is a hands-on company builder across the entire value chain of life sciences investments, from seed to later-stage. The firm actively partners with ambitious entrepreneurs as a lead or cornerstone investor to develop transformative innovations that have the potential to positively impact our collective future.

Founded in 1972, Sofinnova Partners is a deeply established venture capital firm in Europe, with 50 years of experience backing over 500 companies and creating market leaders around the globe. Today, Sofinnova Partners has over 2.5 billion under management. For more information, please visit: sofinnovapartners.com

View source version on businesswire.com: https://www.businesswire.com/news/home/20231025907622/en/

Contacts

Press CEO and Co-FounderJonathan Alexander Brownjonathan@lecole.ai

Follow this link:
Lcole AI, Creator of "Machine Teaching" Technology, Raises 3 Million USD in Seed Funding From Sofinnova Partners - Yahoo Finance