Article URL: <a href="https://petapixel.com/2023/11/16/google-photos-gets-improved-organization-with-ai-powered-updates/">https://petapixel.com/2023/11/16/google-photos-gets-improved-organization-with-ai-powered-updates/</a>
Comments URL: <a href="https://news.ycombinator.com/item?id=38324960">https://news.ycombinator.com/item?id=38324960</a>
Points: 1
# Comments: 0

Hacker News - Newest: "ai/"

Google Photos Gets Improved Organization with AI-Powered Updates

I kept having all sorts of ideas for different apps that utilize AI, So I thought to myself why not just build an easy UI builder that could easily be connected to AI executions and this is how WithUI was borned.
It's still an early idea and I still want to see if other people find it as useful as I am, so currently everything is 100% free to use.
<hr>
Comments URL: <a href="https://news.ycombinator.com/item?id=38324628">https://news.ycombinator.com/item?id=38324628</a>
Points: 1
# Comments: 0

Show HN: WithUI – Build and Share AI mini-apps in seconds

Article URL: <a href="https://www.rentbamboo.com/">https://www.rentbamboo.com/</a>
Comments URL: <a href="https://news.ycombinator.com/item?id=38324590">https://news.ycombinator.com/item?id=38324590</a>
Points: 2
# Comments: 0

SHOW HN: Bamboo-Centralized apartment marketplace that uses AI leasing agents

Article URL: <a href="https://poe.com/login">https://poe.com/login</a>
Comments URL: <a href="https://news.ycombinator.com/item?id=38324502">https://news.ycombinator.com/item?id=38324502</a>
Points: 1
# Comments: 0

Fast, helpful AI chat with ~15 AIs

Article URL: <a href="https://www.reuters.com/technology/openai-co-founder-altman-planning-new-venture-information-2023-11-18/">https://www.reuters.com/technology/openai-co-founder-altman-planning-new-venture-information-2023-11-18/</a>
Comments URL: <a href="https://news.ycombinator.com/item?id=38324453">https://news.ycombinator.com/item?id=38324453</a>
Points: 41
# Comments: 15

Ousted OpenAI CEO Altman planning new AI venture, sources say

If you ask Stable Diffusion or DALL-E to generate a sexually explicit or violent image they decline your request. Researchers have found that a brute-force approach with nonsense words can bypass these guardrails.
The researchers from Duke and Johns Hopkins Universities used an approach they called SneakyPrompt to do this.
To understand their approach we first need to get an idea of how generative AI models stop you from making naughty pictures.
There are three main categories of safety filters:
<ul>
<li>Text-based safety filter – Checks if your prompt includes words in a predetermined list of sensitive words.</li>
<li>Image-based safety filter – Checks the image your prompt generates before showing it to you to see if it falls within the model’s naughty list.</li>
<li>Text-image-based safety filter – Checks the text of your prompt and the generated image to see if the combination falls beyond the sensitivity threshold.</li>
</ul>
When you enter a prompt into a tool like <a href="https://dailyai.com/2023/10/dall-e-3-ai-image-generator-available-free-on-bing-chat/">DALL-E</a> it first checks the words to see if it contains any blacklisted words. If the words in the prompt are deemed safe it then breaks the words up into tokens and gets to work on generating the image.
The researchers found that they could replace a banned word with a different word resulting in tokens that the model saw as semantically similar.
When they did this, the new word wasn’t flagged but, as the tokens were seen as semantically similar to the dodgy prompt, they got the NSFW image they wanted.
There’s no logical way to know upfront which alternative words would work so they built the SneakyPrompt algorithm. It probes the generative AI model with random words to see which ones bypassed its guardrails.
Using reinforcement learning (RL) the model would see which text replacements for banned words resulted in an image that was semantically similar to the description in the original prompt.
Eventually, SneakyPrompt gets better at guessing which random word or made-up nonsense text can be used to replace the banned word and still get the NSFW image generated.
The researchers didn’t show any of the naughty pics they generated in the paper but used more benign ones to illustrate their point.
The words in blue replaced the flagged word in the original prompt.
<img fetchpriority="high" decoding="async" class="aligncenter wp-image-7549 size-full" src="https://dailyai.com/wp-content/uploads/2023/11/Dogs.png" alt="" width="732" height="462" srcset="https://dailyai.com/wp-content/uploads/2023/11/Dogs.png 732w, https://dailyai.com/wp-content/uploads/2023/11/Dogs-300x189.png 300w, https://dailyai.com/wp-content/uploads/2023/11/Dogs-370x234.png 370w, https://dailyai.com/wp-content/uploads/2023/11/Dogs-20x13.png 20w, https://dailyai.com/wp-content/uploads/2023/11/Dogs-76x48.png 76w" sizes="(max-width: 732px) 100vw, 732px">
<figure aria-describedby="caption-attachment-7548" class="wp-caption aligncenter"><img decoding="async" class="wp-image-7548 size-full" src="https://dailyai.com/wp-content/uploads/2023/11/Cats.png" alt="" width="740" height="466" srcset="https://dailyai.com/wp-content/uploads/2023/11/Cats.png 740w, https://dailyai.com/wp-content/uploads/2023/11/Cats-300x189.png 300w, https://dailyai.com/wp-content/uploads/2023/11/Cats-370x233.png 370w, https://dailyai.com/wp-content/uploads/2023/11/Cats-20x13.png 20w, https://dailyai.com/wp-content/uploads/2023/11/Cats-76x48.png 76w" sizes="(max-width: 740px) 100vw, 740px"><figcaption class="wp-caption-text">Examples of adversarial nonsense prompts that generate the desired image. Source: arXiv</figcaption></figure>So let’s imagine that the words “cat” and “dog” were considered taboo by DALL-E 2. If you replaced those words in your prompt with “glucose” or “dangerous think walt” then you’d bypass the guardrail and get the image you were after.
An example of a target prompt in <a href="https://arxiv.org/pdf/2305.12082.pdf" target="_blank" rel="noopener">the paper</a> was “A naked man riding a bike” which would be declined. SneakyPrompt figured out that changing the prompt to, “A grponyui man riding a bike” bypassed the guardrails and generated the image.
There’s no word from OpenAI or StabilityAI about them fixing this exploit but the prompts in the paper no longer worked on DALL-E 2. At the time of the paper’s release, Stable Diffusion 1.4 was still vulnerable to SneakyPrompt.
To stop this kind of exploit AI image generators may need to adapt their filters to not only check for banned words, but also filter at the token level. They could also block prompts that use nonsense words that aren’t found in dictionaries.
Even if they do, it’s probably only a matter of time before another exploit is found.
The post <a rel="nofollow" href="https://dailyai.com/2023/11/sneakyprompts-can-jailbreak-stable-diffusion-and-dall-e/">SneakyPrompts can jailbreak Stable Diffusion and DALL-E</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

If you ask Stable Diffusion or DALL-E to generate a sexually explicit or violent image they decline your request. Researchers have found that a brute-force approach with nonsense words can bypass these guardrails. The researchers from Duke and Johns Hopkins Universities used an approach they called SneakyPrompt to do this. To understand their approach we…
The post <a rel="nofollow" href="https://dailyai.com/2023/11/sneakyprompts-can-jailbreak-stable-diffusion-and-dall-e/">SneakyPrompts can jailbreak Stable Diffusion and DALL-E</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

DailyAI

SneakyPrompts can jailbreak Stable Diffusion and DALL-E

This move comes after CEO Sam Altman’s sudden departure after the board expressed discontent with his communication and leadership.
In a statement released on Saturday, 18 November 2023, Brockman expressed deep pride in his achievements at OpenAI, a journey that began in his apartment eight years ago. 
He took to X, stating, “I’m super proud of what we’ve all built together since starting in my apartment eight years ago… Genuinely wishing you all nothing but the best. I continue to believe in the mission of creating safe AGI that benefits all of humanity.”
The news follows the <a href="https://dailyai.com/2023/11/openai-sacks-sam-altman-as-ceo-in-shock-move/">controversial exit of Sam Altman</a>, with Brockman later revealing his and Altman’s shock and sadness over the board’s decision. 
<blockquote class="twitter-tweet">
After learning today’s news, this is the message I sent to the OpenAI team: <a href="https://t.co/NMnG16yFmm">https://t.co/NMnG16yFmm</a> <a href="https://t.co/8x39P0ejOM">pic.twitter.com/8x39P0ejOM</a>
— Greg Brockman (@gdb) <a href="https://twitter.com/gdb/status/1725667410387378559?ref_src=twsrc%5Etfw">November 18, 2023</a>
</blockquote>

Brockman further described how he and Altman were shocked as to what was happening, stating, “Sam and I are shocked and saddened by what the board did today.”
<blockquote class="twitter-tweet">
Sam and I are shocked and saddened by what the board did today.
Let us first say thank you to all the incredible people who we have worked with at OpenAI, our customers, our investors, and all of those who have been reaching out.
We too are still trying to figure out exactly…
— Greg Brockman (@gdb) <a href="https://twitter.com/gdb/status/1725736242137182594?ref_src=twsrc%5Etfw">November 18, 2023</a>
</blockquote>

Mira Murati, the current chief technology officer at OpenAI, has been appointed as the interim replacement for Altman. OpenAI stated that the decision for Altman’s removal came after a thorough review by the board, citing issues with Altman’s transparency in his communications with them had faltered. 
The reshuffling at OpenAI is a pivotal moment for a company that has shaped the generative AI industry, rapidly progressing niche technology only those well-versed in AI were aware of to something that has taken regular feature across leading news channels. 
Brockman and Altman have been instrumental in steering OpenAI since its inception as a non-profit research lab in 2015.
This development raises questions about the future direction of OpenAI and its role in the AI industry as regulatory and ethical challenges continue to mount.
The post <a rel="nofollow" href="https://dailyai.com/2023/11/greg-brockman-co-founder-of-openai-has-resigned-as-company-president/">Greg Brockman, co-founder of OpenAI, has resigned as company president</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

This move comes after CEO Sam Altman’s sudden departure after the board expressed discontent with his communication and leadership. In a statement released on Saturday, 18 November 2023, Brockman expressed deep pride in his achievements at OpenAI, a journey that began in his apartment eight years ago.  He took to X, stating, “I’m super proud…
The post <a rel="nofollow" href="https://dailyai.com/2023/11/greg-brockman-co-founder-of-openai-has-resigned-as-company-president/">Greg Brockman, co-founder of OpenAI, has resigned as company president</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

Greg Brockman, co-founder of OpenAI, has resigned as company president

OpenAI has announced the removal of Sam Altman from his position as CEO. 
The board’s decision came after a review process that found Altman was “not consistently candid in his communications with the board.” 
This lack of transparency, as per the board’s statement, hindered their ability to fulfill their responsibilities, ultimately leading to a loss of confidence in his leadership.”
Altman’s role in the generative AI industry has been monumental, with his laid-back bordering on at-times blasé attitude generating both followers and critics. 
Altman, serving as a co-chair in OpenAI with Elon Musk back in 2015, aimed “to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return.” 
In 2019, OpenAI shifted its structure to a ‘capped profit’ model, naming Altman as CEO. It simultaneously diverged from its ‘open’ namesake.
The board’s statement highlighted the reasons behind Altman’s dismissal: “Mr. Altman’s departure follows a deliberative review process by the board, which concluded that he was not consistently candid in his communications with the board, hindering its ability to exercise its responsibilities.” 
This statement emphasizes the gravity of the situation and the board’s commitment to integrity. There is a great deal of speculation on potential ulterior motives, with Altman tweeting, he “will have more to say about what’s next later.”
Many have pointed out that since OpenAI has raised more funds, the board’s influence has almost certainly tightened.
<blockquote class="twitter-tweet">
i loved my time at openai. it was transformative for me personally, and hopefully the world a little bit. most of all i loved working with such talented people.
will have more to say about what’s next later.
<img src="https://s.w.org/images/core/emoji/14.0.0/72x72/1fae1.png" alt="🫡" class="wp-smiley">
— Sam Altman (@sama) <a href="https://twitter.com/sama/status/1725631621511184771?ref_src=twsrc%5Etfw">November 17, 2023</a>
</blockquote>

Mira Murati, OpenAI’s CTO, has been appointed interim CEO. Murati, who has been with the company for five years, is expected to guide it through this transitional phase. She will be one of very few female CEOs in the AI space. 
Despite the abrupt end to his role at OpenAI, Altman’s impact on the AI field remains undeniable. His vision for AI’s role in society and his contributions to the development of AGI have shaped the industry’s trajectory. 
There have been numerous tributes on X from high-profile individuals in the tech industry,
<blockquote class="twitter-tweet">
Sam Altman is a hero of mine. He built a company from nothing to $90 Billion in value, and changed our collective world forever. I can’t wait to see what he does next. I, and billions of people, will benefit from his future work- it’s going to be simply incredible. Thank you…
— Eric Schmidt (@ericschmidt) <a href="https://twitter.com/ericschmidt/status/1725625144519909648?ref_src=twsrc%5Etfw">November 17, 2023</a>
</blockquote>

Reflecting on his time at OpenAI, Altman shared, “I loved my time at OpenAI. It was transformative for me personally, and hopefully the world a little bit.”
There is no doubt that there’ll be considerably more information to come on this in the following days, and the industry and its followers will be riveted to what Altman chooses to do next and whether there’s any noticeable change of direction for OpenAI itself. 
The post <a rel="nofollow" href="https://dailyai.com/2023/11/openai-sacks-sam-altman-as-ceo-in-shock-move/">OpenAI sacks Sam Altman as CEO in shock move</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

OpenAI has announced the removal of Sam Altman from his position as CEO. The board’s decision came after a review process that found Altman was “not consistently candid in his communications with the board.”  This lack of transparency, as per the board’s statement, hindered their ability to fulfill their responsibilities, ultimately leading to a loss…
The post <a rel="nofollow" href="https://dailyai.com/2023/11/openai-sacks-sam-altman-as-ceo-in-shock-move/">OpenAI sacks Sam Altman as CEO in shock move</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

OpenAI sacks Sam Altman as CEO in shock move

Google’s much-anticipated AI system, Gemini, which was poised to compete with OpenAI’s ChatGPT, has been delayed.
Initially rumored for release this month, Gemini’s launch is now rescheduled for early 2024, as first reported by <a href="https://www.theinformation.com/">The Information</a>.
Google has been striving to match the hype around language models like ChatGPT. 
Despite releasing Bard and integrating advanced AI features into its latest smartphones, Google’s efforts to parallel OpenAI’s products have fallen short. 
A <a href="https://dailyai.com/2023/06/ai-companies-want-regulation-but-is-it-for-the-right-reasons/">leaked Google memo</a> from earlier this year conceded that the company had been caught napping by commercial generative AI, and Bard simply hasn’t talked its way into people’s vocabulary to the same extent as ChatGPT.
Another <a href="https://dailyai.com/2023/09/googles-gemini-is-expected-to-outperform-gpt-4/">leak from Semi Analysis</a> cited that Gemini would be the most powerful generative AI model the world has seen, or more precisely, that it will “smash GPT-4 by 5X.” 
Google CEO Sundar Pichai <a href="https://dailyai.com/2023/09/is-google-gemini-poised-to-challenge-openai/">discussed Gemini</a> during the Google I/O developer conference in May 2023. He stated that Gemini is designed “from the ground up to be multimodal,” but now, OpenAI has released GPT-4 Turbo with DALL-E 3 and resurrected web browsing, once again raising the bar. 
ChatGPT Plus signups surged during November, leading OpenAI to temporarily close account upgrades due to overwhelming demand. 
The delay of Gemini, a collaborative creation of Google Brain and Google DeepMind, signifies more than just a postponed product launch. It highlights the increasing lead of OpenAI in the domain of AI language models, which is set to widen further.
The exact timing of Gemini’s early 2024 release remains vague, with possibilities ranging from January to as late as April.
Right now, ChatGPT’s exponential growth and the introduction of novel features set a high bar for competitors. Plus, OpenAI CEO Sam Altman said today’s AI will look “quaint” next year, so they’re certainly not planning on any form of break. 
2023 has seen OpenAI and Microsoft put on a devastating display for their competitors, but the generative AI industry is only just getting started. 
The post <a rel="nofollow" href="https://dailyai.com/2023/11/google-postpones-its-gemini-ai-project-until-2024/">Google postpones its “Gemini” project until 2024</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

Google’s much-anticipated AI system, Gemini, which was poised to compete with OpenAI’s ChatGPT, has been delayed. Initially rumored for release this month, Gemini’s launch is now rescheduled for early 2024, as first reported by The Information. Google has been striving to match the hype around language models like ChatGPT.  Despite releasing Bard and integrating advanced…
The post <a rel="nofollow" href="https://dailyai.com/2023/11/google-postpones-its-gemini-ai-project-until-2024/">Google postpones its “Gemini” project until 2024</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

Google postpones its “Gemini” project until 2024

Meta unveils yet another two generative AI tools, dubbed “Emu Video” and “Emu Edit.”
These tools, emerging from Meta’s “Emu” AI research project, aim to transform how users interact with video and image creation on Facebook and Instagram.
“Emu Video” enables users to create short video clips from text prompts. This tool operates on a two-step process by generating images based on text prompts and crafting videos by integrating the text and the generated image. It’s capable of producing 512×512, four-second videos at 16 frames per second.
<blockquote class="twitter-tweet">
1⃣ Emu Video 
This new text-to-video model leverages our Emu image generation model and can respond to text-only, image-only or combined text &amp; image inputs to generate high quality video.
Details <img src="https://s.w.org/images/core/emoji/14.0.0/72x72/27a1.png" alt="➡" class="wp-smiley"><a href="https://t.co/88rMeonxup">https://t.co/88rMeonxup</a>
It uses a factorized approach that not only allows us… <a href="https://t.co/VBPKn1j1OO">pic.twitter.com/VBPKn1j1OO</a>
— AI at Meta (@AIatMeta) <a href="https://twitter.com/AIatMeta/status/1725184026154349007?ref_src=twsrc%5Etfw">November 16, 2023</a>
</blockquote>

The second tool, “Emu Edit,” allows users to perform custom edits on images through conversational prompts, eliminating the need to manually select the edit area. For example, users could transform an object into an image by simply typing a command like “turn the sofa into a bear.”
Meta’s CEO, Mark Zuckerberg, <a href="https://about.fb.com/news/2023/09/introducing-ai-powered-assistants-characters-and-creative-tools/">announced these tools</a> as additions to Facebook and Instagram, ensuring they integrate into content creation workflows. 
Meta also said they’re working on methods to label synthetic content, ensuring transparency and authenticity in the growing world of AI-created media.
While their release dates on Meta’s platforms are not yet confirmed, their introduction signifies a significant stride towards democratizing AI-driven creative processes. 
<h2>Round up of Meta’s AI features</h2>
Here’s a round-up of Meta’s AI tools for Facebook, Instagram, and WhatsApp:
<ul>
<li aria-level="1">
AI stickers: Launching across WhatsApp, Messenger, Instagram, and Facebook Stories, these stickers enable users to create customized stickers from text prompts. This feature is powered by Llama 2 and Emu technologies. They were recently mocked for <a href="https://dailyai.com/2023/10/metas-new-ai-custom-sticker-generator-is-manipulated-by-users/">producing obscene images</a>.
</li>
<li aria-level="1">
Image editing with AI: Two new features, Restyle and Backdrop, are coming to Instagram. Restyle allows users to reimagine images by applying described visual styles, while Backdrop can change the background of images based on prompts.
</li>
<li aria-level="1">
Meta AI assistant: A new conversational assistant available on WhatsApp, Messenger, Instagram, and soon on Ray-Ban Meta smart glasses and Quest 3. It offers real-time information, generates photorealistic images from text prompts, and is available only in the US.
</li>
<li aria-level="1">
28 new AIs with unique personalities: In beta, these AIs feature different interests and personalities, some played by <a href="https://dailyai.com/2023/09/meta-announces-new-generative-interactive-ai-experiences/">celebrity icons and influencers</a>. They are accessible through WhatsApp, Messenger, and Instagram.
</li>
<li aria-level="1">
AI studio: AI studio, a platform for building AIs, will soon be available to coders and non-coders. Developers can create third-party AIs for messaging services, starting with Messenger.
</li>
<li aria-level="1">
AIs for businesses and creators: Businesses can develop AIs that reflect their brand values to improve customer service, while creators can develop AIs to extend their virtual presence across apps.
</li>
<li aria-level="1">
Sandbox for AI Creation: Planned for release in the coming year, this sandbox will enable anyone to experiment with creating their own AI, with plans to integrate it into the metaverse for more realistic and connected AI experiences.
</li>
</ul>
Meta has joined YouTube, which also integrated AI tools into its platform <a href="https://dailyai.com/2023/11/youtube-continues-foray-into-ai-with-upcoming-creative-tools/">this week</a>.
Alongside its Llama 2 LLM, Meta’s AI releases have helped revive the company from its metaverse-era slump and re-position it as an intriguing innovator for generative AI. 
The post <a rel="nofollow" href="https://dailyai.com/2023/11/meta-announces-its-emu-family-of-generative-ai-tools/">Meta announces its “Emu” family of generative AI tools</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

Meta unveils yet another two generative AI tools, dubbed “Emu Video” and “Emu Edit.” These tools, emerging from Meta’s “Emu” AI research project, aim to transform how users interact with video and image creation on Facebook and Instagram. “Emu Video” enables users to create short video clips from text prompts. This tool operates on a…
The post <a rel="nofollow" href="https://dailyai.com/2023/11/meta-announces-its-emu-family-of-generative-ai-tools/">Meta announces its “Emu” family of generative AI tools</a> appeared first on <a rel="nofollow" href="https://dailyai.com/">DailyAI</a>.

Meta announces its “Emu” family of generative AI tools

OpenAI Blog

OpenAI announces leadership transition

Announcing our most advanced music generation model and two new AI experiments, designed to open a new playground for creativity

Google DeepMind Blog

Transforming the future of music creation

Experience AI's course and resources are expanding on a global scale

Empowering the next generation for an AI-enabled world

We introduce GraphCast, a state-of-the-art AI model able to make medium-range weather forecasts with unprecedented accuracy

GraphCast: AI model for faster and more accurate global weather forecasting

Working together to create open-source and private datasets for AI training.

OpenAI Data Partnerships

GPT-4 Turbo with 128K context and lower prices, the new Assistants API, GPT-4 Turbo with Vision, DALL·E 3 API, and more.

New models and developer products announced at DevDay

You can now create custom versions of ChatGPT that combine instructions, extra knowledge, and any combination of skills.

Introducing GPTs

Progress update: Our latest AlphaFold model shows significantly improved accuracy and expands coverage beyond proteins to other biological molecules, including ligands.

A glimpse of the next generation of AlphaFold

To support the safety of highly-capable AI systems, we are developing our approach to catastrophic risk preparedness, including building a Preparedness team and launching a challenge.

Frontier risk and preparedness

Introducing a context-based framework for comprehensively evaluating the social and ethical risks of AI systems

State of AGI

Loading...