Todd Underwood

Many folks wonder about the role and impact of software in fighting climate change… Conversations often go to “well the thing we need most is (low carbon) steel in the ground,” referring to the need for hardware tech for difficult to decarbonize industries. And while this is very true, it’s not innovation that’s needed to accelerate low carbon technologies, it’s money. But money alone doesn’t get us there either. Funding and financing only flow in a major way when the analysis, due diligence, and underwriting are done at speed, scale, and importantly with confidence and certainty. So we return to our starting question. What role can software play? Software can unlock information which unlocks financing which enables steel to go in the ground which then lets us decarbonize hard to abate industries. Enter Halcyon. A mashup of seasoned operators from big tech and energy industry veterans building an AI-assisted search and information platform that helps people navigate the energy transition and decarbonization. I’ve had the joy of working with co-founders Bruce Falck Nat Bullard J L. and Alexander Huras as well as their PMF team Vicky Homan Alyssa Thavorides Alex Klaessig Ted Schaller Jeff Fisher and am so excited about their launch and their future. Check out the Axios piece below and go get on their waitlist at www.halcyon.eco. https://lnkd.in/gsyg7Mcb

32

5 Comments

Cool article on how Fugaku, Japan’s fastest supercomputer, has gone virtual on AWS. “For these scientific outcomes to proliferate and create an actual impact on society, just building this 160,000-node machine is far from sufficient. The value of Fugaku is not only the hardware, but also the software solutions including the applications. We’ve injected this supercomputing environment directly into the cloud to extend applications and have much larger impact than in a single machine.” #Supercomputing #AWS #cloud Link to the article in the first comment.

23

3 Comments

Setting a vision is just the start. Think about the bold moves needed to make it happen. Sundar, Google's CEO, set a daring vision to tackle climate change. That's where the real work begins. Executing a vision requires bold and tough decisions. Google is buying nuclear energy to feed its AI data centers. Other big tech companies are following this lead. This is a big kudos to Sundar and his leadership team. They are executing on a sustainability vision. Leadership is about taking action. And Sundar is showing the way.

27

12 Comments

When you align precinct-level 2020 election data from the Harvard Dataverse with place data from the Overture Maps Foundation (10M POIs from Meta and Microsoft) to examine the relationship between local political leanings and the types of businesses and venues present in neighborhoods across the U.S.

2

As Trust and Safety professionals from across the country gather at #TrustCon, and as Chuck Schumer signals he will bring the controversial Kids Online Safety Act to the Senate floor for a vote this week, I hope this legislation is being debated and ultimately opposed by the field. Although all of us in this space work with the goal to protect children, teens, and everyone online, queer magazine Them reminds us that #KOSA gives right-wing attorneys general new power to censor the internet in their states. As I wrote in Tech Policy Press a few weeks ago, safety doesn’t come from censorship, it comes from empowerment. “Healthy humans are empowered to make choices, form secure connections, and set boundaries–without having to navigate recommendation systems that degrade our defenses with every new video and introduction. Teens, families, and all people are desperate for tools that help us discover, connect, share, and disconnect with balance and choice.” I hope T&S professionals are discussing how their products can empower their users to make choices, set boundaries, and find the information they need, and are signaling to their employers that KOSA promotes censorship, not safety. Check out my piece in TPP and the review of KOSA in Them, below:  TPP: https://lnkd.in/giczi625 Them: https://lnkd.in/gvf3WRCt

8

Should advanced AI be regulated or made freely available to everyone without limits? A debate that took place using Voxiberate produced some very interesting results... Read the post to learn more...

15

Read up on the first part of our approach at the Anti-Censorship Technical Committee (ACTC), as we discuss the problems of centralization and show how ContentSafe can be used to spread content across platforms. It's a lot less likely that content will be removed from more than a dozen platforms than just one or two! https://lnkd.in/ezAq4JBi

Protecting Adolescents in the Digital Age Social media platforms have revolutionized how we connect, but at what cost? In the next part of a recent talk I gave, I explore how platforms like #facebook use the "hook model" to capture attention, creating cycles of dependency that particularly impact adolescents. From the emotional toll of #fomo (fear of missing out) to the risks of inadequate #privacy and #security settings, our kids face challenges that demand our attention. I also discuss the duty of care tech companies must uphold to protect young users, including implementing secure defaults, age verification, and parental controls that can’t be bypassed. But it’s not just about #technology, parenting alignment and environmental adjustments play a huge role in shaping safer digital experiences for the next generation. Let’s work together to make the internet a safer, healthier place for adolescents. What steps are you taking to protect young people online? #cybersafety #adolescentwellness #digitalparenting #onlinesafety #privacymatters

4

Interested in turning transcripts into Knowledge Graphs that you can ask complex questions across multiple Congressmen, political issues, and Committees? In this WhyHow.AI case study, we show how Congressional Committee Transcripts are turned into Temporal Knowledge Graphs that you can rely on for the purposes of RAG and analytics. We show what real Question & Answers are against this system, and what type of business outcomes you can achieve with this system. As far as we are aware, the combination of steps here is a novel Knowledge Graph implementation, and incorporates the use of unstructured.io and Senzing for data parsing. Here, we showcase all the inputs into the Process, including a step by step of the data transformation process, breakdown of the 20 dev hours involved to build the entire system end-to-end, schemas used, Questions & Responses, and a link to the Graph created. Chris Rec Thomas Smoker https://lnkd.in/eGavCPch

52

5 Comments

Interesting news from Anthropic. Anthropic is partnering with Palantir and Amazon Web Services (AWS) to bring the Claude family of AI models to U.S. intelligence and defense agencies. Earlier this summer, select Claude models made their way to AWS GovCloud, AWS' dedicated service for U.S. government cloud workloads. This marks a strategic push by Anthropic to expand their footprint in the public sector. For cybersecurity professionals, the availability of managed private cloud AI models like Claude can be a game-changer. These models can significantly enhance SOC operations, enabling teams to analyze threats, respond to incidents, and develop cyber threat defense applications at unprecedented speed and efficiency. This partnership is a major step forward for applying cutting-edge AI to the defense and intelligence sectors. What are your thoughts on the role of AI in the future of cybersecurity, or even more interesting, How are Attackers using AI to create new threats? #Anthropic #AWS #Palantir

2

The most hypocritical company on the planet today is Microsoft (perhaps ever). Here is Vasu Jakkal accurately warning about the risks from LLMs many of us have long warned about, that her boss Satya enabled, and is promoting and accelerating as fast as possible, with the apparent full assistance of POTUS and his cabinet (esp Commerce). The reason for this record hypocrisy is of course perverse incentives between Microsoft's old cash cows like Office now hosted on Azure, combined with opportunity in LLMs, and their rapidly growing $20 billion security business. The monopoly applications enables if not causes risks, while the security business benefits. Customers: you lose, they win. It's like big tobacco during its peak owning the largest healthcare company, with no oversight or enforcement (voluntary...). Reportedly driven by Bill Gates, Satya and MSFT's board are obviously hoping AI will allow MSFT to break through the scale ceiling and continue to grow, while their "top diplomat" Brad Smith leads the effort attempting to keep the USG (and other governments around the world) as a political partner rather than execute their first responsibility in safety and security. All I can say is enjoy government capture and corruption while it lasts as there will surely be hell to pay during future administrations.

55

18 Comments

Interested in creating a unique, conversational-style podcast from your own content in 5 minutes? Take a look at Rich Heimann's latest post, where Google's NotebookLM generated a conversational podcast video (including jokes) in just 5 minutes after being fed our new 60,000-word book. I know it's not perfect, but it's pretty freaking spectacular! Imagine the impact this could have on your business. If you're unaware, NotebookLM turns books and documents into an AI-powered NPR-style podcast plus a study guide, FAQ, timeline, and a surprisingly accurate chat experience—all in one. These results, on the heels of OpenAI's reasoning results with o1 are serious steps towards creating lasting enterprise value. AJ Eads Macy Dennis Mitch Wells Jeremy. Mcdonald David Starr

3

There are many LLM security papers out there that do something odd and get potentially harmful content out. But rarely is there any clue if this is against policy - ie., is unintended. Reviewers are left to judge in the dark what policy a given LLM was built on. It's a mess. Values and norms come from everywhere and vary everywhere, and without explicit policy known, it's impossible to judge whether that model's policy has been circumvented, or the difficulty/significance of doing so.  For example, if an LLM has no policy, all outputs are intended. It might be unsafe and a generally poor move to build and distribute it, but being able to get bad things out of a model with no policy is NOT a security problem.

97

13 Comments

In the spirit of RTFM, just read up the Anthropic Responsible Scaling Policy (RSP). Kudos to the Trust & Safety team and security folks Anthropic for all the good work. My thoughts on things which I really liked and things which I would like to see. https://lnkd.in/gts2nkTc

18

Meta just released Llama 3, their most powerful open-source AI model yet. But that's just the beginning... Meta's chief AI scientist, Yann LeCun, revealed that a more capable successor with 400B+ parameters is already in training. Some key takeaways: • Llama 3 comes in 8B and 70B parameter versions, outperforming previous open models • Meta believes open sourcing-accelerates AI progress and will ultimately prevail • Llama 3 used 7X more training data than Llama 2 • A new Meta AI assistant powered by Llama 3 is coming to WhatsApp, Instagram, and more • But some worry the increasingly powerful open models could be misused or become hostile • Meta's licensing also restricts what researchers can do with the models The AI arms race is heating up as #Meta, #OpenAI, #Google, and others push for artificial general intelligence. It will be interesting to see the new applications, research, and 30,000+ variants from the Llama 3 release. One thing is clear - open-source AI is progressing at breakneck speed. #AI #ML #generativeai #LLM #OpenSource #Llama #emergingtechnologies

1

1 Comment

So the 'great grid upgrade' is all about pylons, right? Wrong. We need to move away from the widely held perception that pylons are the only way to upgrade the grid. There are, in fact, at least three better ways, all of which are cheaper: - HVDC UNDERGROUNDING (& SUB-SEA) Germany has chosen this method instead of overhead lines. It has lower lifetime costs, is far less destructive than building pylons and is popular. - INTEGRATION OFFSHORE. It might surprise you to know that there is no strategy or coordination of offshore wind farm connections. Each developer is told where to connect, and that could be a long way in land e.g Norwich. We could make developers' lives easier by offering them connection points at offshore platforms. It also saves money and reduces overall infrastructure if you coordinate and integrate into an offshore grid. It is popular, too. (37,000 people have signed our petition). - UPGRADE THE EXISTING GRID TO THE MAX! In the US there is a new requirement to look at upgrading the existing grid before building new infrastructure. There are all sorts of new conductors which transmit far more power using the same towers (pylons). And in East Anglia we have a disused substation & disused power lines from Bradwell that need to be brought back into use. Wayleaves in place. Key landowner willing. Here's how to make it a GREAT grid upgrade that's cheaper, better & popular. Let's raise the level of conversation from pylons. They should be the LAST resort.: https://lnkd.in/eRMS7ECB

16

4 Comments

“Anonymized, pseudonymized, de-identified or encrypted data should be excluded,” Business Roundtable wrote, “if processes prevent re-identification.” I love anonymized data exemptions. I think they’re the best way to balance AI innovation with privacy. I think they’re underutilized and under-appreciated. I think more data should be collected, anonymized, and shared with more people to drive progress. But this is preposterous. This comment was submitted by a tech industry trade group in response to recent Justice Department rulemaking that limits the sale of personal information to foreign adversaries like Russia and China. The deliberate mashing together of anonymous data with pseudonymous and de-identified data is reckless, given the stakes. Pseudonymous data is *designed to be joined with other data*. That’s why every modern data protection law treats pseudonymous data as personal information. It’s why the FTC rejected Avast’s the claim that joinable, pseudonymous data could be treated as "anonymized." Anonymous data cannot be re-identified or linked with other data. It carries qualitatively and quantitatively different risks than pseudonymous and de-identified data. There are a bunch of considerations for whether an anonymous data exemption makes sense when the likely recipients are hostile nation-states. This may be the rare case where the theoretical re-identification risks inherent in any anonymized data are too harmful to tolerate. But an exemption for pseudonymous and de-identified data is unconscionable. NOTE: Advertising companies provide a valuable service that funds the accessibility of the modern internet and are unfairly villainized in too many privacy conversations. But when they add “anonymized” data into their arguments, it’s intentionally misleading and deserves condemnation. They will not be selling anonymous data, because anonymous data can’t be used for targeted advertising. That term is added to the list to take the edge off selling re-identifiable data for advertising. Thanks to Justin Sherman, who does incredible work in this space, for highlighting this in Politico!  

45

2 Comments