•
February 15, 2024
•
7 mins
Hello, Niuralogists!
Step into this week's edition as we navigate the dynamic realm of artificial intelligence to present you with the most recent breakthroughs. Our primary focus is to dissect the implications of these updates on various aspects of our lives, ranging from workplaces and businesses to policies and individual experiences. In this release, we'll reveal intriguing advancements, including a closer look at OpenAI's exploration of enhanced memory features in ChatGPT and the rising role of AI in Super Bowl advertisements.
For in-depth insights, keep on reading...
OpenAI has initiated a groundbreaking experiment to enhance the memory capabilities of ChatGPT, catering to select users. This initiative aims to empower ChatGPT to retain essential details from previous conversations, thereby facilitating more insightful and personalized responses in subsequent interactions. Users are granted control over the information stored, with the ability to delete specific details, clear the entire memory, or disable the feature altogether. ChatGPT's capacity to discern preferences, contexts, instructions, and other pertinent information allows it to continually improve and recall details across conversations. Moreover, GPTs can harness distinct memory capabilities when activated by their developers. The experiment is currently being introduced to a small percentage of both free and users, with a wider rollout anticipated in the future. This development is significant as features like memory represent crucial advancements towards achieving more natural and understanding chatbot interactions, striking a balance between personalization and user control. The prospect of tailored AI assistants is steadily approaching, promising a more nuanced and user-centric conversational experience.
Artificial Intelligence took center stage in this year's Super Bowl commercials, offering a glimpse into both its advantages and potential absurdities in the era of artificial intelligence. Microsoft featured its Copilot AI assistant in a prominent spot. At the same time, a Google Pixel ad showcased the new AI 'Guided Frame' feature designed to assist blind or low-vision users in capturing photos. Illumination took a humorous approach to AI in a promotion for the upcoming 'Despicable Me 4' film, where Minions hilariously generated subpar video content. A controversial HeGetsUs ad, featuring people washing feet, faced criticism amid speculation that the images were AI-generated, though Ad Age clarified they were actual photographs. BodyArmor's 'Field of Fake' mocked AI-generated content, emphasizing the superiority of 'real sweeteners' over 'artificial' alternatives through low-quality AI videos. Etsy's commercial highlighted its Gift Mode, an AI recommendation tool aiding users in curating gift ideas. The Super Bowl AI ads signify a significant mainstream moment for the technology, reflecting the diverse sentiments and ongoing conflicts surrounding AI across the country.
Salesforce-owned Slack has introduced its highly anticipated generative AI features, collectively known as Slack AI, to enhance user communication on the platform. The AI capabilities include AI-powered search, channel recaps, and thread summaries. The new AI-powered search feature allows users to ask questions conversationally and receive concise, generated answers based on relevant Slack messages, overcoming the limitations of the current search tool. Channel Recaps summarize messages in inaccessible channels, and Thread Summaries provide detailed overviews of message threads with a single click. Slack AI addresses privacy and security concerns, with a commitment to responsible and ethical AI product development. Slack plans to add more AI-powered features in the future, making Slack AI available as a paid add-on for Slack Enterprise plans.
In recent legal developments, OpenAI has experienced a positive turn as a judge dismissed most of the copyright infringement lawsuit brought by comedian Sarah Silverman and co-plaintiffs against the AI company. The lawsuit accused OpenAI of violating copyrights by training its AI models, GPT-3.5 and GPT-4, on books without consent. The judge ruled against four out of the six counts, inviting the plaintiffs to amend and re-file by March 13. The only count allowed to proceed alleges that OpenAI engaged in "unfair" business practices by training its AI models on the books. This decision is generally seen as a victory for OpenAI, indicating a potential precedent for similar lawsuits involving AI companies and creatives. The judge emphasized that the plaintiffs did not provide sufficient evidence of direct copying or infringement in ChatGPT's responses. The case remains open, pending the plaintiffs' amended claims and further legal proceedings.
Google has officially rebranded its AI chatbot Bard as Gemini, unveiling significant upgrades that include the introduction of a new mobile app and access to its highly anticipated Ultra model. The Gemini Advanced, featuring Ultra 1.0, is touted as Google's "largest and most capable" AI model, competing with GPT-4 on performance benchmarks. Access to the Advanced model is part of a $19.99/month Google One AI Premium subscription plan. The Gemini mobile app is now available on Android and iOS in the U.S., offering enhanced on-device assistance through typing, voice, and image inputs. The Android app supports various existing Google Assistant functionalities, such as making calls, controlling smart devices, and accessing information from other Google apps. Notably, images generated with Gemini will now include an invisible digital watermark using a tool called 'SynthID.' The launch of Gemini in Canada, with support for both English and French, marks the completion of the transition from Bard. The move intensifies the competition in the large language model (LLM) space, with Ultra challenging ChatGPT for the top model status, prompting anticipation for OpenAI's response with a potential release of GPT-4.5 or GPT-5.
📬 Receive our amazing posts straight to your inbox. Get the latest news, company insights, and Niural updates.
Engineers at the University of Groningen have examined how insects navigate to develop energy-efficient robots. Insects, despite having brains the size of a pinhead, exhibit remarkable navigational skills. They effectively avoid obstacles and navigate through small openings, a feat that poses a challenge given their limited brain power. Researchers aim to leverage insights from insect behavior for energy-efficient computing. A recent development includes a robot, guided by a model based on the neuronal activity of insects, navigating through environments, and showcasing the efficiency of insect-like behavior. The study suggests that emulating the hardwired efficiency of insects in computer systems could lead to more energy-efficient applications.
AI is reshaping cybersecurity through extended detection and response (XDR), as organizations transition from Endpoint Detection and Response (EDR). XDR's ability to consolidate functions and limit data movement aligns with CISO priorities, supported by AI-driven analysis of behaviors for real-time threat detection. Leading XDR vendors, like CrowdStrike and Palo Alto Networks, leverage AI for consolidation and superior outcomes. This year sees a surge in security tech stack consolidation, with Gartner projecting a significant uptick in XDR adoption. AI plays a crucial role in closing identity and endpoint security gaps, addressing challenges posed by new identities, and providing real-time threat detection, behavioral analysis, and reduction of false positives. The financial impact of AI on XDR platforms not only eases short-term budgetary concerns but positions them for long-term success in predicting intrusions and identifying breaches. AI is becoming standard in XDR technology maturity.
Stability AI, renowned for its Stable Diffusion text-to-image generative AI, introduces a new model named Stable Cascade. The company aims to explore innovative image-generation approaches that surpass the efficiency of existing Stable Diffusion models. Unlike its predecessor, Stable Cascade adopts the Würstchen architecture, employing a unique latent diffusion technique for a more detailed and compact semantic image representation. The three-stage modular architecture of Stable Cascade, consisting of Stages A, B, and C, enhances training efficiency and customization. This approach, accompanied by Direct Preference Optimization, promises improved image quality, faster inference times, and superior text generation within images. Stable Cascade, available for non-commercial use, showcases advanced capabilities like generating image variations and supporting image-to-image translations.
👋 Heyday transforms your content into shareable, searchable insights
📚 Study Summary GPT summarizes lectures into concise study notes for better understanding
🚀 MagicAds is human-like UGC ads which effortlessly creates with AI
👗 The Stylist GPT unlocks your style potential with expert outfit selection and shopping assistance
🔎 Kondor is an AI assistant with the power of sight
.png)