Jorge Alcántara

Spot manipulative tactics. Check out these five videos, they are short and can help keep you thinking clearly.

1

Cleanlab Introduces the Trustworthy Language Model (TLM) that Addresses the Primary Challenge to Enterprise Adoption of LLMs: Unreliable Outputs and Hallucinations https://zurl.co/yqQO #ai #enterpriseai #genai #llm

1

AI Disaster Looming? California Fights Back! A contentious AI-focused bill disturbs the tech scene in California. This isn't just legislative drama; it's a balance between innovation and possible disaster. Imagine an AI-gone-wild scenario. California is ensuring it never occurs. But there's retaliation. The tech industry—California's economic powerhouse—isn't silent. Their rebuttals are sharp and persuasive. Want to know more? Dive in here: https://lnkd.in/efTjPYiU Your insight could be crucial in this innovation vs. regulation battle. Share your thoughts! Follow me for more AI Tips, Tools, News and Fun #ai #aidisaster

1

"United Independents, the group behind the Independent National Convention set for Denver in September, has launched a new software platform for indie candidates called Electify.us, which came out of a partnership with incubator Wolfram Labs — the first tool to emerge from its program." I am so proud to share that Chief Product Officer Dana Cichon has launched the first political tech platform to come out of our accelerator program. What are you waiting for? We have the formula for breaking into the digital campaign space. Are you are the next success story? Read more in Campaigns & Elections -->> https://lnkd.in/eHhqa-Xr #techaccelerator #startups #incubator #VC #SaaS #software #political #campaigns

15

Also, interesting distinctions drawn here between a fuckup, a shitshow and a clusterfuck. https://lnkd.in/eqi_3jst

2

1 Comment

Should you use the Oxford comma? The Oxford comma, also known as the serial comma, is a small punctuation mark with a mighty role in technical documentation. Its primary function is to clarify the meaning of sentences by indicating a separation between list items.

83

10 Comments

Very insightful work by Anthropic’s interpretability team. And an amazing paper, with outstanding writing and figures. The idea is very simple: interpret LLMs by leveraging sparse autoencoders as surrogate models of the MLP of transformer blocks, which allow one to disambiguate the superposition of features captured by a single neuron. A simple idea, but a very careful and complex execution, as it is often the case in our line of work. The paper goes into many details and provide a large array of insights, although the gist of the implementation remains obfuscated due to the closed source nature of Claude. Too bad, because this is the kind of work that we need to better understand and eventually trust LLMs. This is demonstrated by the authors in the section ‘Influence on Behavior’, where they show that clamping some features to either high or low value during inference is “remarkably effective at modifying model outputs in specific, interpretable ways”. Hopefully this kind of work is going to be replicated and generalized to open-weights models, such that we have new ways to steer their behavior. https://lnkd.in/eVym7f_f #interpretability #xai #explainableai #steerableai #anthropic #claude #anthropic

3

WEB3 SDLC for A zk-SNARK (Zero-Knowledge Succinct Non-Interactive Argument of Knowledge) language refers to the programming languages or frameworks used to write and work with zk-SNARKs, which are cryptographic proofs enabling one party to prove to another that they know a value, without revealing the value itself and with minimal communication between the parties. These languages facilitate the creation of such proofs by providing the tools and abstractions needed to define and manipulate the cryptographic constructs involved. https://lnkd.in/geSiG-9D

3

1 Comment

Last week, I spoke about AI and regulations at an event at the U.S. Capitol attended by legislative and business leaders. I’m encouraged by the progress the open source community has made fending off regulations that would have stifled innovation. But opponents of open source are continuing to shift their arguments, with the latest worries centering on open source's impact on national security. I hope we’ll all keep protecting open source! I’m encouraged by the progress the U.S. federal government has made getting a realistic grasp of AI’s risks. To be clear, guardrails are needed. But they should be applied to AI applications, not to general-purpose AI technology. Nonetheless, some companies are eager to limit open source, possibly to protect the value of massive investments they’ve made in proprietary models and to deter competitors. It has been fascinating to watch their arguments change over time. For instance, about 12 months ago, the Center For AI Safety’s “Statement on AI Risk” warned AI could cause human extinction and stoked fears of AI taking over. This alarmed leaders in Washington. But many people in AI pointed out that this dystopian science-fiction scenario had little basis in reality. About six months later, when I testified at the U.S. Senate’s AI Insight forum, legislators no longer worried much about an AI takeover. Then the opponents of open source shifted gears. Their leading argument shifted to the risk of AI helping to create bioweapons. Soon afterward, OpenAI and RAND showed that current AI does not significantly increase the ability of malefactors to build bioweapons. This fear of AI-enabled bioweapons has diminished. To be sure, that bad actors could use bioweapons — with or without AI — remains a topic of great international concern. The latest argument for blocking open source AI has shifted to national security. AI is useful for both economic competition and warfare, and open source opponents say the U.S. should make sure its adversaries don’t have access to the latest foundation models. While I don’t want authoritarian governments to use AI, particularly to wage unjust wars, the LLM cat is out of the bag, and authoritarian countries will fill the vacuum if democratic nations limit access. When, some day, a child asks an AI system questions about democracy, the role of a free press, or an independent judiciary, I would like the AI to reflect democratic values rather those of authoritarian leaders. I came away from Washington optimistic about the progress we’ve made defending open source. Looking beyond the U.S. federal government, there are many jurisdictions. Unfortunately, arguments for regulations that would stifle AI continue to proliferate. But talking to regulators does have an impact. If you have a chance to talk to a regulator, I hope you’ll do what you can to help governments better understand AI. [Original text in The Batch (with links): https://lnkd.in/g5SZ6ypa ]

1,614

58 Comments

if you know, you know (ai) factories with your own personal librarian tim berners-lee https://lnkd.in/eH6afx-b. while at cern https://lnkd.in/e8t_Fpdw for purposes of communicating with other scientists, created hypertext. ooops. today tim b-l is conflicted. the internet has become walled gardens. kinda like this https://lnkd.in/eTGWsjHV, but everywhere. this image is a reminder. a reminder that when we read these books it's upon what we have learned/read perviously. do you have a love of reading. why? have you wondered why? what do you get from reading? do you know why? who cares if you've read anything at all. have you wondered why? there are other walls. those walls are to exclude books. previously we had this happen over and over and over again https://lnkd.in/eDqh9maA to control culture books have burned. prior to this there was this...https://lnkd.in/eNpm79wc rather than burning there is another way. just leave them alone. let them simply rot. do not open them. indeed create a narrative that there is nothing there. as much as we want to believe that Google has organised the world's information. the facts are there are many many many books that remain closed. their spines uncracked. so there are ways to alter, shape culture. so be it with (ai). the data is the foundation. for example if we only had in existence say aesop's fables https://lnkd.in/eB_A7Kbm this would be the way we communicate. so by scraping the internet there is a bias. and there is much that is not included. and there are many who have created a walled garden. and so. and so. and sow. creativity. this moment is a rare moment to create, invent, design, build. optimitically. read. venture. (a)dventure. < . >

1

A bill pending in the California Senate that would essentially make many start-ups, open source projects as well as some research illegal in California and force it to move to other states. We need to stop this nonsense. Usually I don't post about politics as issues are often nuanced, complex and hard to fully address in a short post. This is an exception. The bill has zero benefit for anyone, except maybe very large tech companies that can use it to crush stat-ups. And it doesn't mitigate any potential downside of AI. However it would kill many open source efforts. It would stop some research. And it would force start-ups to do their model development in other states (or other countries). If you can, please support the effort to stop this. https://stopsb1047.com/

51

5 Comments

If you're using LLMs in production you need to be running systematic evals. Here's a quick tutorial on how to get started using DeepEval, an open-source eval framework, to run evals in just minutes. https://lnkd.in/gvEifRxT

3

A handful of well-funded and established companies are seeking to dominate the AI landscape and suppress the existence of open source initiatives. The only reason anyone with a laptop and a decent internet connection can do simple things like: - Learn AI through resources like Fast.ai, Andrew Ng's courses, and Andrej Karpathy's work - Run language models on their preferred platform, be it a personal computer or a cloud instance for free or at a low cost is due to the power of open source. We must not allow a select few companies to dictate what the world should do or determine what is best for each nation. When these companies raise concerns about national security, it is essential to remember that the internet is not the product of a single nation. It is a global collaboration, built and contributed to every day by people from all corners of the world. Todays AI would not exist if not for our collective collaboration.

57

I hear a lot of logical explanations why certain ideas simply make sense. But you can't theorize your way into building a great product. Sure, you may get lucky, but luck is not reliable. To consistency build high impact products you need to practice product discovery.

27

1 Comment

Our paper, "QGen: On the Ability to Generalize in Quantization Aware Training," (https://lnkd.in/gXXvS_9Q) is now available on Arxiv. This research delved into the generalization capabilities of quantized models, a relatively unexplored domain, examining them through the lens of sharpness in the loss landscape during optimization. I found the experiments and mathematical proofs particularly fulfilling. We're hopeful for acceptance at the conference to which it's been submitted. 🤞

47

Can we cut to the chase and just say that the exploitation of human data by tech companies aligned with 'generative' AI is now bordering on EVIL? It knows no boundary. It must be free, everything, without permission, consent, compensation, approval, removal. How long till Facebook is cataloguing the same for its own AI models off of imagery of your children? Correlating pictures across the last 18 years of uploads? ---- Wired: "AI Tools Are Secretly Training on Real Images of Children A popular AI training dataset is “stealing and weaponizing” the faces of Brazilian children without their knowledge or consent, human rights activists claim." Over 170 images and personal details of children from Brazil have been repurposed by an open-source dataset without their knowledge or consent, and used to train AI." Wired: https://lnkd.in/gN8B-pFr ---- ArsTechnica: "AI trained on photos from kids’ entire childhood without their consent. Kids "easily traceable" from photos used to train AI models, advocates warn." (ASHLEY BELANGER) (6/10/2024) Photos of Brazilian kids—sometimes spanning their entire childhood—have been used without their consent to power AI tools, including popular image generators like Stable Diffusion, Human Rights Watch (HRW) warned on Monday. This act poses urgent privacy risks to kids and seems to increase risks of non-consensual AI-generated images bearing their likenesses, HRW's report said. An HRW researcher, Hye Jung Han, helped expose the problem. She analyzed "less than 0.0001 percent" of LAION-5B, a dataset built from Common Crawl snapshots of the public web. The dataset does not contain the actual photos but includes image-text pairs derived from 5.85 billion images and captions posted online since 2008. Among those images linked in the dataset, Han found 170 photos of children from at least 10 Brazilian states. These were mostly family photos uploaded to personal and parenting blogs most Internet surfers wouldn't easily stumble upon, "as well as stills from YouTube videos with small view counts, seemingly uploaded to be shared with family and friends," Wired reported. LAION, the German nonprofit that created the dataset, has worked with HRW to remove the links to the children's images in the dataset. Removing the links also does not remove the images from the public web, where they can still be referenced and used in other AI datasets, particularly those relying on Common Crawl, LAION's spokesperson, Nate Tyler, told Ars. Han told Ars that "Common Crawl should stop scraping children’s personal data, given the privacy risks involved and the potential for new forms of misuse." According to HRW's analysis, many of the Brazilian children's identities were "easily traceable," due to children's names and locations being included in image captions that were processed when building the LAION dataset." ArsTechnica: https://lnkd.in/g3e5VskN #ai #aiethics #fakeai #aievil #geni #degenai #degenartiveai #degenerativeai #dobeevil

19

6 Comments

Cracking the code of life faster than ever before. For researchers, diving into the world of high-performance computers can feel like navigating a labyrinth. Yet, the payoff is monumental, especially when it comes to predicting protein structures. Imagine a future where understanding the building blocks of life is as straightforward as running a program. We're inching closer to that reality, thanks to the fusion of supercomputing and AI. The complexity of these computational behemoths has often been a barrier. But as we peel back the layers, we uncover the potential to revolutionize how we approach biological mysteries. This isn't just about faster results. It's about opening doors to medical breakthroughs, personalized medicine, and solutions to problems we haven't even encountered yet. As daunting as the technical challenges may seem, the journey towards demystifying the digital and biological interface is one worth embarking on. For those standing at the crossroads of technology and biology, the message is clear: Embrace the complexity, for the rewards could redefine our understanding of life itself. Read more about how supercomputing is shaping the future of protein structure prediction here: [Insert URL] Let's dive into this new era, where biology meets computing power, head first. Check this out: https://lnkd.in/ePxHVCJ9

2