Home Blog Page 529

12 Days of OpenAI

Emerson Blake
-
0
12 Days of OpenAI

What are the ’12 days of OpenAI’?

OpenAI has announced a campaign called the “12 days of OpenAI”, where the company will host 12 days of live streams and release “a bunch of new things, big and small” starting from December 5.

What has been dropped so far?

Thursday, December 5:

OpenAI started with a bang, unveiling two major upgrades to its chatbot: a new tier of ChatGPT subscription, ChatGPT Pro, and the full version of the company’s o1 model.

Full version of o1:

* Will be better for all kinds of prompts, beyond math and science
* Will make major mistakes about 34% less often than o1-preview, while thinking about 50% faster
* Rolls out today, replacing o1-preview to all ChatGPT Plus and now Pro users
* Lets users input images, as seen in the demo, to provide multi-modal reasoning (reasoning on both text and images)

ChatGPT Pro:

* Is meant for ChatGPT Plus superusers, granting them unlimited access to the best OpenAI has to offer, including unlimited access to OpenAI o1-mini, GPT-4o, and Advanced Mode
* Features o1 pro mode, which uses more computing to reason through the hardest science and math problems
* Costs $200 per month

Where can you access the live stream?

The live streams are held on the OpenAI website, and posted to its YouTube channel immediately after. To make access easier, OpenAI will also post a link to the live stream on its X account 10 minutes before it starts, which will be at approximately 10 a.m. PT/1 p.m. ET daily.

What can you expect?

The releases remain a surprise, but many anticipate that Sora, OpenAI’s video model initially announced last February, will be launched as part of one of the bigger drops. Since that first announcement, the model has been available to a select group of red teamers and testers and was leaked last week by some testers over grievances about “unpaid labor,” according to reports.

Other rumored releases include a new, fuller version of the company’s o1 LLM with more advanced reasoning capabilities, and a Santa voice for OpenAI’s Advanced Voice Mode, per code spotted by users only a couple of weeks ago under the codename “Straw”.

Conclusion

The “12 days of OpenAI” campaign promises to be an exciting event, with new releases and demos every day. From the full version of o1 to ChatGPT Pro, there’s a lot to look forward to. Make sure to tune in to the live streams on the OpenAI website and YouTube channel to stay up-to-date on all the latest developments.

FAQs

Q: What is the “12 days of OpenAI” campaign?
A: The “12 days of OpenAI” is a campaign where OpenAI will host 12 days of live streams and release “a bunch of new things, big and small” starting from December 5.

Q: What has been dropped so far?
A: OpenAI has released two major upgrades to its chatbot: a new tier of ChatGPT subscription, ChatGPT Pro, and the full version of the company’s o1 model.

Q: How can I access the live stream?
A: The live streams are held on the OpenAI website, and posted to its YouTube channel immediately after. To make access easier, OpenAI will also post a link to the live stream on its X account 10 minutes before it starts.

Q: What can I expect from the releases?
A: The releases remain a surprise, but many anticipate that Sora, OpenAI’s video model, and other new features will be launched as part of the campaign.

Atlas

Parker Lennox
-
0
Atlas

Introduction

Neural networks can learn to classify images more accurately than any system humans directly design. This raises a natural question: What have these networks learned that allows them to classify images so well?

Feature Visualization

Feature visualization is a thread of research that tries to answer this question by letting us "see through the eyes" of the network. It began with research into visualizing individual neurons and trying to determine what they respond to. Because neurons don’t work in isolation, this led to applying feature visualization to simple combinations of neurons. However, there was still a problem – what combinations of neurons should we be studying? A natural answer (foreshadowed by work on model inversion) is to visualize activations, the combination of neurons firing in response to a particular input.

Activation Atlases

These approaches are exciting because they can make the hidden layers of networks comprehensible. These layers are the heart of how neural networks outperform more traditional approaches to machine learning, and historically, we’ve had little understanding of what happens in them. Feature visualization addresses this by connecting hidden layers back to the input, making them meaningful.

Unfortunately, visualizing activations has a major weakness – it is limited to seeing only how the network sees a single input. Because of this, it doesn’t give us a big-picture view of the network. When what we want is a map of an entire forest, inspecting one tree at a time will not suffice.

Aggregating Multiple Images

Activation grids show how the network sees a single image, but what if we want to see more? What if we want to understand how it reacts to millions of images? Of course, we could look at individual activation grids for those images one by one. But looking at millions of examples doesn’t scale, and human brains aren’t good at comparing lots of examples without structure. In the same way that we need a tool like a histogram to understand millions of numbers, we need a way to aggregate and organize activations if we want to see meaningful patterns in millions of them.

Activation Atlases in Practice

To create activation atlases, we collect activations from one million images, randomly select one spatial activation per image, and avoid the edges due to boundary effects. This gives us one million activation vectors – spatial consistency is necessary for making atlases zoomable. Thus, we preferred this method over clustering in this article, but finding trade-offs between these techniques remains an open question.

Discussion and Review

In this article, we introduce activation atlases to the quiver of techniques that give a global view of a network. By combining feature visualization with the approach of showing feature visualizations of averaged activations, we can get the advantages of each in one view – a global map seen through the eyes of the network.

Author Contributions

Shan Carter wrote the majority of the article and performed most of the experiments. Zan Armstrong helped with the interactive diagrams and the writing. Ludwig Schubert provided technical help throughout and performed the numerical analysis of the manual patches. Ian Johnson provided inspiration for the original idea and advice throughout. Chris Olah provided essential technical contributions and substantial writing contributions throughout.

Acknowledgments

Thanks to Kevin Quealy and Sam Greydanus for substantial editing help. Thanks to Colin Raffel, Arvind Satyanarayan, Alexander Mordvintsev, and Nick Cammarata for additional feedback during development. We’re also very grateful to Phillip Isola for stepping in as acting Distill editor for this article, and to our reviewers who took time to give us feedback, significantly improving our paper.

FAQs

Q: What are activation atlases?
A: Activation atlases are a new way to visualize the internal workings of neural networks, showing how they respond to different inputs and patterns in the data.

Q: How do they work?
A: By collecting and organizing the activations from millions of images, we can create a zoomable map of how the network sees the world, showing how different parts of the network respond to different types of images.

Q: What are the limitations of activation atlases?
A: While activation atlases can reveal high-level misunderstandings in a model, they are limited to the distribution of the data used to train the network, and may not generalize to new data.

Q: How can I use activation atlases?
A: By studying the activation atlases, you can gain a deeper understanding of how your own models work, and identify potential biases or limitations in your training data.

The Genius Tech Behind Crunchy Game Design

Phil
-
0
The Genius Tech Behind Crunchy Game Design

Article

CB: What core experience were you aiming to deliver to players, and how did the dynamic camera play into that?

From the very beginning of the project, we set out to reimagine what ‘cinematic action’ can mean in games. It’s common for cinematic games to be criticised for relying too heavily on cutscenes and quick-time events (QTEs). We wanted to challenge that convention by pushing the boundaries wherever possible.

Our philosophy is simple: if something exciting is happening on screen, we aim to make it part of the gameplay rather than a passive cutscene. For example, if characters are fighting on a train rooftop, we want that moment to be fully playable, not just a cutscene. A version of this scene was featured in one of our early trailers, showcasing our approach.

As directors of an ‘action movie within a game’, we embrace bold camera choices, including the use of pre-set cameras in certain scenes. Many of these decisions are inspired by iconic moments from classic action films, but the most important feature we are committed to implementing is what we call an ‘intellectual camera system’.

What have been the biggest challenges in implementing this feature?

The main challenge lies in finding the balance between convenience and cinematography. For the camera work, we’ve collaborated with a director of photography from the film industry, who has taught us many valuable techniques – small but impactful tricks that we’ve either implemented or keep in mind.

For instance, we’ve incorporated into the game the idea that a real-life camera is a physical object with its own inertia, and that action scenes in movies are often shot at eye level or lower, because it’s difficult for a cameraman to raise the camera above their head. We aim to capture that realistic camera movement, understanding that real movie cameras introduce subtle "errors" that bring life to the shot. These include constant movement and a breathing parallax effect, which we plan to mocap.

Another key consideration is that in movies, the camera typically stays close to the character, but in beat ’em up games, players often prefer a more distant view to see everything happening on screen. Ultimately, it all comes down to selecting the right variables: What field of view (FOV) do we use? How do we adjust it dynamically? How long do we take control away from the player during finishing moves? What angle best showcases a character’s action, like how we show Redline delivering a kick to her opponent (we prefer a slight angle that highlights the action)?

Were there any specific games, films, or other forms of media that influenced the way you approached the use of camera movement in SPINE?

Absolutely! Nothing beats the camerawork in Uncharted 4, especially during the prison fight scene. In fact, that very fight was one of the key inspirations behind our ‘intellectual camera’ feature in SPINE.

I also think that, for example, Sifu does a great job of capturing the essence of kung fu movies. I particularly appreciate some of the finishers in that game.

Big-budget games from PlayStation Studios also pay close attention to camerawork, with Insomniac’s work on Marvel’s Spider-Man being a prime example. They use some impressive techniques, like adjusting the FOV/zoom and applying post-effects, which really enhance the experience. I’ve heard they decided against implementing ‘paired takedowns’ of enemies – well, good news for SPINE players: we’ll have them!

How did you ensure that the camera enhanced the experience without disorienting or frustrating players?

We understand that some players prefer complete control over the in-game camera. That’s why we allow players to override our systems at any time. The moment you move the right stick, the camera controls are entirely in your hands, and you can adjust the angles however you like. However, our goal is for players to feel so engaged with our camera system that they won’t feel the need to take control.

Our internal play tests have shown that 90% of players are satisfied with the camerawork and don’t experience frustration (excluding some bugs, which are expected at this stage of SPINE development). All the gameplay scenes featured in our gameplay trailers were recorded without touching the right stick.

How did using Unreal Engine 5 enable the camera to be implemented?

When we first started working on the camera system, we only had a rough idea of the features we wanted. The rest has come through continuous iteration. New ideas emerge, we test them, refine them, customise them, and sometimes even completely redesign them.

Different situations in the game require different approaches – for example, the camera behaviour for a boss fight is different from how it works with a group of five regular enemies, and there are various preset cameras for specific moments.

As a result, we don’t rely heavily on the default Unreal Engine 5 camera algorithms. However, the Blueprint system has been incredibly helpful in supporting our iterative development process.

Not only are Blueprints easy to learn, making them accessible for different team members to experiment with and customise, but they also enable the creation of complex and sophisticated systems. Overall, we can confidently say that adopting Unreal Engine 5 has been a highly successful decision for our studio.

Conclusion

Nekki is pushing the boundaries of game development with its innovative dynamic camera system in SPINE. By incorporating Unreal Engine 5, the team has been able to create a unique and engaging experience that immerses players in the world of the game.

The dynamic camera system is designed to make every player action feel unique, forceful, and crunchy, drawing inspiration from iconic action films and games. The system allows for a range of camera movements and angles, ensuring that players are fully immersed in the action on screen.

FAQs

Q: What was the biggest challenge in implementing the dynamic camera system in SPINE?
A: Finding the balance between convenience and cinematography.

Q: What games or films influenced the way you approached the use of camera movement in SPINE?
A: Uncharted 4, Sifu, and big-budget games from PlayStation Studios.

Q: How do you ensure that the camera enhances the experience without disorienting or frustrating players?
A: By allowing players to override the camera system at any time and providing a range of camera movements and angles.

Q: How did using Unreal Engine 5 enable the camera to be implemented?
A: The Blueprint system in Unreal Engine 5 enabled the creation of complex and sophisticated camera systems, allowing for continuous iteration and refinement.

AI Discovers Cancer Signs Missed by Doctors

Taylor Marlowe
-
0
AI Discovers Cancer Signs Missed by Doctors

AI Tool Proves Capable of Detecting Cancer Signs Overlooked by Human Radiologists

An AI tool, called Mia, has been piloted alongside NHS clinicians in the UK and analyzed the mammograms of over 10,000 women. The results are impressive, with the AI successfully flagging all cases of breast cancer and identifying 11 additional cases that human radiologists missed.

How the AI Tool Works

The AI tool was trained on a dataset of over 6,000 previous breast cancer cases, allowing it to learn the subtle patterns and imaging biomarkers associated with malignant tumors. When evaluated on new cases, the AI correctly predicted the presence of cancer with 81.6% accuracy and correctly ruled it out 72.9% of the time.

Breast Cancer: A Global Concern

Breast cancer is the most common cancer in women worldwide, with over two million new cases diagnosed annually. While survival rates have improved with earlier detection and better treatments, many patients still experience severe side effects like lymphedema after surgery and radiotherapy.

Future Development

Researchers are now developing the AI system further to predict a patient’s risk of such side effects up to three years after treatment. This could allow doctors to personalize care with alternative treatments or additional supportive measures for high-risk patients. The research team plans to enroll 780 breast cancer patients in a clinical trial called Pre-Act to prospectively validate the AI risk prediction model over a two-year follow-up period.

Conclusion

The potential for AI in healthcare is vast, and this study demonstrates its ability to improve early detection and treatment of breast cancer. With further development, AI could revolutionize the way we approach cancer diagnosis and treatment, leading to better outcomes and improved patient care.

Frequently Asked Questions

Q: How many women were included in the study?
A: Over 10,000 women were included in the study.

Q: How accurate was the AI system in detecting breast cancer?
A: The AI system correctly predicted the presence of cancer with 81.6% accuracy and correctly ruled it out 72.9% of the time.

Q: What is the goal of the Pre-Act clinical trial?
A: The goal of the Pre-Act clinical trial is to prospectively validate the AI risk prediction model over a two-year follow-up period.

Q: What is the potential of AI in breast cancer diagnosis and treatment?
A: AI has the potential to improve early detection and treatment of breast cancer, leading to better outcomes and improved patient care.

David Sacks is Trump’s AI and crypto ‘czar’

Avery Winslow
-
0
David Sacks is Trump’s AI and crypto ‘czar’

Guiding the Future of American Competitiveness

In this important role, David will guide policy for the Administration in Artificial Intelligence and Cryptocurrency, two areas critical to the future of American competitiveness.

Leading the Way in Artificial Intelligence

David will focus on making America the clear global leader in artificial intelligence. This will be achieved by developing and implementing policies that foster innovation and investment in AI research and development, as well as ensuring that the technology is used responsibly and ethically.

Ensuring Free Speech Online

David will work to safeguard free speech online by promoting a legal framework that protects individuals’ and companies’ ability to share information and ideas without fear of censorship or retaliation from big tech companies.

Stemming Big Tech Bias and Censorship

In addition to safeguarding free speech, David will also work to steer the Administration away from big tech bias and censorship. This includes ensuring that online platforms do not discriminate against certain types of content or users and that they are transparent and accountable in their decision-making processes.

Creating a Clarifying Legal Framework for Crypto

David will work to create a legal framework for the crypto industry that provides clarity and stability for stakeholders. This includes developing regulatory guidelines that allow the crypto industry to thrive in the U.S. while protecting consumers and preventing illicit activity.

Conclusion

In conclusion, David’s role in the Administration will be critical to shaping the future of American competitiveness in artificial intelligence, cryptocurrency, and online freedom. His focus on safeguarding free speech, addressing big tech bias and censorship, and creating a clear legal framework for crypto will have a lasting impact on the country’s economic and social landscape.

FAQs

  • What will be the focus of David’s role in the Administration?

    David’s role will focus on guiding policy for the Administration in Artificial Intelligence and Cryptocurrency, two areas critical to the future of American competitiveness.

  • Will David work to make America the global leader in Artificial Intelligence?

    Yes, David will focus on making America the clear global leader in artificial intelligence through the development and implementation of policies that foster innovation and investment in AI research and development.

  • Will David work to promote free speech online?

    Yes, David will work to safeguard free speech online by promoting a legal framework that protects individuals’ and companies’ ability to share information and ideas without fear of censorship or retaliation from big tech companies.

  • Will David work to address big tech bias and censorship?

    Yes, David will work to steer the Administration away from big tech bias and censorship by ensuring that online platforms do not discriminate against certain types of content or users and that they are transparent and accountable in their decision-making processes.

  • Will David create a clear legal framework for the crypto industry?

    Yes, David will work to create a legal framework for the crypto industry that provides clarity and stability for stakeholders, while protecting consumers and preventing illicit activity.

Broadcom Reverses Controversial Plan to Cull VMware Migrations

Emerson Blake
-
0
Broadcom Reverses Controversial Plan to Cull VMware Migrations

Broadcom Shifts Strategy to Save VMware Business

Broadcom to Work with Top 500 VMware Customers, Channel Partners to Gain

Broadcom will no longer take VMware’s biggest 2,000 customers directly. Instead, it will work with VMware’s 500 biggest customers, giving channel partners the opportunity to participate in deals and provide additional value for VMware customers. This reversal is being viewed as an effort from Broadcom to discourage migrations from VMware, but there’s skepticism around how much impact it will truly have.

Customer Laments and Migrations

Various customers have lamented the changes that succeeded Broadcom buying VMware about a year ago. Controversial moves have included ending perpetual license sales, bundling VMware products into a smaller number of SKUs, and ending VMware’s channel partner program. These changes have led some firms to consider reducing their business with VMware.

Recent Examples of Migrations

This week, for example, UK-headquartered cloud operator Beeks Group said that a 1,000 percent increase in VMware costs led to it moving most of its 20,000-plus virtual machines to OpenNebula. And numerous customers that Ars Technica has spoken with in the last year are seriously researching or planning total or partial VMware migrations.

Broadcom’s New Strategy

Now, Broadcom is looking to save some business by incorporating channel partners into deals that it previously ushered them out of. In January, CRN reported that Broadcom took over more than 2,000 of VMware’s biggest accounts, circumventing partners and confusing some partners and customers. In a March earnings call, CEO Hock Tan said Broadcom would focus on upselling those accounts. As The Register reported today, Broadcom recently announced that it will only work directly with the top 500 VMware accounts.

Broadcom’s Statement

In a statement, a Broadcom spokesperson said:

Broadcom continues to work on behalf of our partners to create new value in capturing the market opportunity for private cloud. Most recently, we announced a program that is currently in development to offer qualified VCF customers a 15 percent professional service entitlement of their annual contract value to access partner-delivered or Broadcom professional services. This will help customers improve both time to value and ROI. Broadcom does not have an official, static number of direct strategic accounts. The number of customers with whom we work directly changes over time.

Canalys’ Analysis

At Canalys’ APAC Forum event today, Canalys chief analyst Alastair Edwards said that "Broadcom recognizes that its best defense against possible migrations is making sure customers implement its full private cloud bundles and see strong return on investment. Broadcom sees giving 1,500 big users back to partners as the way to make that happen, and is even giving its channel 15 percent of the value of deals they win to fund professional services so that VMware software is quickly made operational," per The Register.

Conclusion

Broadcom’s new strategy may help to mitigate some of the damage caused by its previous actions, but it remains to be seen how effective it will be in preventing migrations from VMware. The company’s decision to work with top 500 customers and channel partners may help to create new value and improve the return on investment for customers, but it is unclear whether this will be enough to stem the tide of customer defections.

FAQs

Q: What is Broadcom’s new strategy?
A: Broadcom will work with the top 500 VMware customers and channel partners to provide additional value and improve the return on investment for customers.

Q: Why is Broadcom making this change?
A: Broadcom is making this change to discourage migrations from VMware and to create new value for customers by working with channel partners.

Q: How will this affect channel partners?
A: Channel partners will have the opportunity to participate in deals and provide additional value for VMware customers, with 15 percent of the value of deals they win funding professional services.

Q: Will this strategy be effective?
A: It is unclear whether this strategy will be effective in preventing migrations from VMware, but it may help to create new value and improve the return on investment for customers.

Canva Revolutionized Graphic Design

Avery Winslow
-
0
Canva Revolutionized Graphic Design

Canva’s Journey to Success

From Creativity to Productivity

Right from the start, we had this Venn diagram: On one side is creativity, and on the other side is productivity. And you might guess, right in the center is Canva. We really believe that people on the productivity side actually want to be more creative, and that people on the creative side want to be more productive. And so we really found that to be the sweet spot—it was a huge gap in the market that we saw right in the early days, and it’s where we’re continuing to invest very heavily.

How Canva Uses Canva

Extremely extensively, for literally everything. Our engineers do their engineering docs in Canva, we do all-hands, I do all of my product mock-ups in it. I’ve used it for decision decks and vision decks and onboarding and hiring and recruitment—name something, we’re using Canva for it very extensively.

Valuation and Market Shift

Your peak valuation was $40 billion in 2021. A year later, this was cut to $26 billion. What happened? I think it was purely the macro shift in the market. During that time, Canva has continued to grow rapidly, both on revenue and active users. We’ve been profitable for seven years as well, so even though the market [switched to caring] more about profitability, we were fortunately already on that trend. Markets are going to value different things over time, and markets are going to be frothy and then not frothy. We are just always caring about building a strong, enduring company with good foundations that serves our community. So it’s not a particular bother what’s happening out there in the market.

Pledging 30% of Equity to Doing Good

You’ve pledged 30 percent of Canva—the majority of your and Obrecht’s equity—to doing good in the world. What does that mean to you? It seems completely absurd that we have the prosperity that we do across the globe, and there are people that still don’t have basic human needs being met. The first step that we’ve taken is partnering with GiveDirectly, where we give money directly to people who are living in extreme poverty. [Canva has so far donated a total of $30 million to people living in poverty in Malawi.] I love the empowerment that gives them to be able to spend the money on their community, on their family, on their basic human needs—sending their kids to school, getting a roof over their head. We have an extremely long way to go, but we’re really excited that we’ve started that process.

Reaching 1 Billion Users

You aim to reach 1 billion users. What’s the plan to get there? When we set that as a goal a number of years ago, it seemed completely ridiculous, but over the years, it’s becoming less ridiculous. We need about one in five internet users in every country to reach a billion. Now in the Philippines it’s one in six internet users, and in Australia it’s one in eight internet users. In Spain, it’s one in 11. In the USA, it’s one in 12. So at 200 million now, we’re a fifth of the way towards the billion number, and if we can continue to grow as rapidly as we have been, we’ll hopefully get there.

IPO Plans

Any plans to IPO? It’s definitely something on the horizon.

Conclusion

Canva has come a long way since its early days, and it’s clear that the company is committed to continuing to grow and make a positive impact on the world. With its focus on creativity and productivity, and its pledge to do good in the world, Canva is poised for continued success in the future.

FAQs

Q: What is Canva’s plan to reach 1 billion users?

A: Canva needs about one in five internet users in every country to reach a billion. Currently, it has reached a fifth of the way towards this goal, with 200 million users.

Q: What is Canva’s valuation?

A: Canva’s peak valuation was $40 billion in 2021, and it was cut to $26 billion the following year due to a macro shift in the market.

Q: How does Canva use Canva?

A: Canva uses Canva extensively for a wide range of purposes, including engineering docs, all-hands, product mock-ups, decision decks, vision decks, onboarding, hiring, and recruitment.

Q: What is Canva’s plan for doing good in the world?

A: Canva has pledged 30 percent of its equity to doing good in the world, and has started by partnering with GiveDirectly to give money directly to people living in extreme poverty.

ChatGPT o1: World’s Smartest Language Model

Jordan Waverly
-
0
ChatGPT o1: World’s Smartest Language Model

OpenAI Rolls Out World’s Smartest Language Model, Introduces New Pro Tier

OpenAI ChatGPT o1 Model

Sam Altman announced on his Twitter account that the new AI model is now live and available in ChatGPT, and will be arriving to the API soon.

He tweeted: "o1, the smartest model in the world. Smarter, faster, and more features (e.g. multimodality) than o1-preview. Live in chatgpt now, coming to api soon."

Screenshot Of ChatGPT 01 Model Availability

[Image: Screenshot of ChatGPT 01 Model Availability]

ChatGPT 01 Limits Uploads To Images

The new model does not allow uploads of text, PDF files, or CSV files. The only filetypes allowed to be uploaded are image files. It is not known whether this is a feature or a bug, but a user will have to downgrade to the 4o model if they want ChatGPT to analyze anything other than an image file.

Initial Confusion About $200 Pricing

Altman’s tweet was interpreted as announcing a price increase of the $20 ChatGPT tier to $200. One of the first responses praised the high benchmark performance of the new o1 model but lamented that ChatGPT was now out of reach for many users.

Sam Altman immediately responded: "Everything you share here is available in the $20 tier!"

ChatGPT Pro Mode $200/Month

ChatGPT Pro Mode is a new tier that has more "thinking power" than the standard version of o1, increasing its reliability. Answers in Pro mode take longer to generate, displaying a progress bar and triggering an in-app notification if the user navigates to a different conversation.

Conclusion

OpenAI’s new o1 model is a significant development in the field of language processing, and the introduction of the Pro tier offers users even more advanced capabilities. With its increased processing power and multimodal capabilities, the o1 model is poised to revolutionize the way we interact with AI.

FAQs

Q: What is the new OpenAI model called?
A: The new model is called o1.

Q: What is ChatGPT Pro Mode?
A: ChatGPT Pro Mode is a new tier that has more "thinking power" than the standard version of o1, increasing its reliability and producing more accurate and comprehensive responses.

Q: Is the new model available in the standard ChatGPT tier?
A: No, the new model is only available in the new Pro tier.

Q: Is the new model available in the API?
A: The new model will be arriving to the API soon.

Q: Can I upload files other than images to the new model?
A: No, the new model only allows uploads of image files.

Q: How much does the new Pro tier cost?
A: The new Pro tier costs $200 per month.

Amazon Taps Automated Reasoning to Safeguard Critical AI Systems

Parker Lennox
-
0
Amazon Taps Automated Reasoning to Safeguard Critical AI Systems

Amazon’s Aggressive AI Adoption: Automated Reasoning to the Rescue

Amazon is implementing AI aggressively across its business in a bid to improve operational efficiency, delight customers, and ultimately make money. But adopting probabilistic systems that don’t always behave as expected and are prone to hallucinations also comes with risks. To help minimize AI-related risks, Amazon and its AWS subsidiary are turning to a time-tested but little-known technique dubbed automated reasoning.

What is Automated Reasoning?

Automated reasoning is a field of computer science designed to provide greater certainty about the behavior of complex systems. At its core, automated reasoning gives adopters strong assurances, based on logic and mathematics, that a system will do what it was designed to do.

How Does Automated Reasoning Work?

Automated reasoning is a rules-based approach that uses mathematical logic to prove the correctness of systems and design systems in architecture code. Traditionally, these techniques were used in things like aerospace, where it’s critical to get systems correct.

Amazon’s Use of Automated Reasoning

Since 2016, Neha Rungta, the director of applied science at AWS, has been using her expertise to help AWS improve the security of its services. Her AWS resume includes two products, including IAM Access Analyzer, which is used to analyze Amazon IAM (Identity and Access Management) and its 2 billion requests per second, and Amazon S3 Block Access.

Automated Reasoning Checks

At re:Invent on Tuesday, AWS announced that it’s using automated reasoning with Amazon Bedrock, its service for training and running foundation models, including large language models (LLMs) and image models. The company said the service, dubbed Automated Reasoning Checks, is the "the first and only generative AI safeguard that helps prevent factual errors due to hallucinations using logically accurate and verifiable reasoning."

Why Isn’t Automated Reasoning More Widely Used?

The reason, Rungta said, is that automated reasoning comes with a cost. It’s not so much the computational costs of running the automated reasoning model, but the cost in developing and testing it. Adopters require not only expertise in this small branch of the AI field, but also in the domain for which automated reasoning is being applied.

Conclusion

Amazon is looking to become a leader as the GenAI era takes off. The company has more than 1,000 AI projects internally, according to Amazon founder Jeff Bezos, who spoke at the New York Times’s DealBook conference this week. As we begin the agentic AI era, we’ll see that different AI agents have different jobs. It’s likely that we’ll see some AI agents that function as supervisors of worker agents, and these supervisory agents may be developed with automated reasoning capabilities.

FAQs

Q: What is automated reasoning?
A: Automated reasoning is a field of computer science designed to provide greater certainty about the behavior of complex systems.

Q: How does automated reasoning work?
A: Automated reasoning is a rules-based approach that uses mathematical logic to prove the correctness of systems and design systems in architecture code.

Q: Why isn’t automated reasoning more widely used?
A: The reason is that automated reasoning comes with a cost. It’s not so much the computational costs of running the automated reasoning model, but the cost in developing and testing it.

Q: What is the future of automated reasoning?
A: As some of these LLMs get smaller and better tuned to specific domains, the easier and less costly it will be to apply automated reasoning techniques to them.

2025 Predictions: Humanoids and AI Agents

Emerson Blake
-
0
2025 Predictions: Humanoids and AI Agents

The Rise of Generative AI and Its Impact on Industries

The Year of Accelerated AI Adoption

The adoption of generative AI and large language models is rippling through nearly every industry, as incumbents and new entrants reimagine products and services to generate an estimated $1.3 trillion in revenue by 2032.

Agentic AI on the Horizon

The next big thing on the horizon is agentic AI, a form of autonomous or "reasoning" AI that requires using diverse language models, sophisticated retrieval-augmented generation stacks, and advanced data architectures.

Industry Insights

Inference Drives the AI Charge

"As AI models grow in size and complexity, the demand for efficient inference solutions will increase," said Ian Buck, Vice President of Hyperscale and HPC. "The rise of generative AI has transformed inference from simple recognition of the query and response to complex information generation – including summarizing from multiple sources and large language models such as OpenAI o1 and Llama 450B – which dramatically increases computational demands. Through new hardware innovations, coupled with continuous software improvements, performance will increase and total cost of ownership is expected to shrink by 5x or more."

Accelerate Everything

"With GPUs becoming more widely adopted, industries will look to accelerate everything, from planning to production. New architectures will add to that virtuous cycle, delivering cost efficiencies and an order of magnitude higher compute performance with each generation," said Buck.

Quantum Computing

"Quantum computing will make significant strides as researchers focus on supercomputing and simulation to solve the greatest challenges to the nascent field: errors," said Buck. "Qubits, the basic unit of information in quantum computing, are susceptible to noise, becoming unstable after performing only thousands of operations. This prevents today’s quantum hardware from solving useful problems. In 2025, expect to see the quantum computing community move toward challenging, but crucial, quantum error correction techniques. Error correction requires quick, low-latency calculations. Also, expect to see quantum hardware that’s physically colocated within supercomputers, supported by specialized infrastructure."

AI in Management

"AI will also play a crucial role in managing these complex quantum systems, optimizing error correction, and enhancing overall quantum hardware performance," said Buck.

Putting a Face to AI

"AI will become more familiar to use, emotionally responsive, and marked by greater creativity and diversity," said Bryan Catanzaro, Vice President of Applied Deep Learning Research. "The first generative AI models that drew pictures struggled with simple tasks like drawing teeth. Rapid advances in AI are making image and video outputs much more photorealistic, while AI-generated voices are losing that robotic feel."

Rethinking Industry Infrastructure and Urban Planning

"Nations and industries will begin examining how AI automates various aspects of the economy to maintain the current standard of living, even as the global population shrinks," said Catanzaro. "These efforts could help with sustainability and climate change. For instance, the agriculture industry will begin investing in autonomous robots that can clean fields and remove pests and weeds mechanically. This will reduce the need for pesticides and herbicides, keeping the planet healthier and freeing up human capital for other meaningful contributions. Expect to see new thinking in urban planning offices to account for autonomous vehicles and improve traffic management."

AI-Orchestrated Enterprises

"Enterprises are set to have a slew of AI agents, which are semiautonomous, trained models that work across internal networks to help with customer service, human resources, data security, and more," said Kari Briski, Vice President of Generative AI Software. "To maximize these efficiencies, expect to see a rise in AI orchestrators that work across numerous agents to seamlessly route human inquiries and interpret collective results to recommend and take actions for users."

Predicting Unpredictability

"Expect to see more models that can learn in the everyday world, helping digital humans, robots, and even autonomous cars understand chaotic and sometimes unpredictable situations, using very complex skills with little human intervention," said Sanja Fidler, Vice President of AI Research.

Getting Real

"Fidelity and realism is coming to generative AI across the graphics and simulation pipeline, leading to hyperrealistic games, AI-generated movies, and digital humans," said Fidler.

The Startup Workforce

"If you haven’t heard much about prompt engineers or AI personality designers, you will in 2025," said Nader Khalil, Director of Developer Technology. "As businesses embrace AI to increase productivity, expect to see new categories of essential workers for both startups and enterprises that blend new and existing skills."

Understanding Employee Efficiency

"Startups incorporating AI into their practices increasingly will add revenue per employee (RPE) to their lexicon when talking to investors and business partners," said Andrew Feng, Vice President of GPU Software. "Instead of a ‘growth at all costs’ mentality, AI supplementation of the workforce will allow startup owners to home in on how hiring each new employee helps everyone else in the business generate more revenue."

Conclusion

Generative AI is transforming industries, and 2025 is expected to be a pivotal year for the technology. As AI models continue to advance, we can expect to see even more widespread adoption, increased productivity, and new opportunities for businesses and individuals alike.

FAQs

Q: What is generative AI?
A: Generative AI is a type of artificial intelligence that can generate new content, such as text, images, or music, rather than just analyzing or processing existing data.

Q: What are the benefits of generative AI?
A: Generative AI can revolutionize industries by enabling the creation of new products, services, and experiences that were previously impossible. It can also improve customer service, streamline processes, and reduce costs.

Q: What are some potential challenges of generative AI?
A: Some potential challenges of generative AI include ensuring data quality, managing bias, and addressing ethical considerations. Additionally, there may be concerns about job displacement and the concentration of wealth and power in the hands of a few individuals or companies.

Q: What are some potential applications of generative AI?
A: Some potential applications of generative AI include healthcare, finance, education, marketing, and entertainment. It can also be used in industries such as manufacturing, transportation, and energy to improve efficiency and reduce costs.

1...528529530...740Page 529 of 740

Welcome to MachineDaily.AI, your go-to destination for breaking news, in-depth analysis, and the latest developments in the world of artificial intelligence (AI). Our mission is to provide you with timely and insightful content that explores AI’s vast potential across industries and its growing impact on society. From cutting-edge innovations to ethical debates, MachineDaily.AI is committed to covering AI in all its forms, making complex topics accessible to experts, enthusiasts, and curious minds alike.

Latest

Categories

Useful Links

Our Newsletter

Subscribe Us To Receive Our Latest News Directly In Your Inbox!

© Copyright 2024. All Right Reserved By Machine Daily.