Рубрика: AI

OpenAI just released Sora, a video-generation AI tool that creates hyper-realistic videos from prompts. While Sora is impressive, it is sure to add to the mounting concerns on potential job losses for VFX artists owing to the rise of generative AI.

Last year, the International Alliance of Theatrical Stage Employees, the overarching organisation pioneering the inaugural VFX union, advised Hollywood to exercise caution in utilising AI within the industry.

Similarly, a major VFX studio owned by Netflix last year hired AI experts, raising a few eyebrows. The studio was looking for individuals with a comprehensive understanding of prompt engineering, experience in neural network image-generating space, and knowledge of the stable diffusion ecosystem, among other things.

A recent report titled ‘FUTURE UNSCRIPTED: The Impact of Generative Artificial Intelligence on Entertainment Industry Jobs‘ underscores that generative AI is poised to instigate a substantial shift from conventional techniques to innovative processes. This transition is expected to recalibrate the demand for labour and capital across the entertainment industries.

However, industry experts AIM spoke to hold a rather different view.

“When ‘The Lion King’ was originally created through hand-drawn animation 35 years ago, the production process significantly differed from the recent remake with Disney and Jon Favreau. In the modern iteration, we employed virtual reality (VR) to finalise the entire film.

Subsequently, we immersed ourselves in animating it with photorealistic precision. The current production of ‘The Lion King’ utilised an array of advanced tools and technologies, involving a larger team compared to the production of the original film several years ago,” Biren Ghose, country head at Technicolor India, told AIM.

GenAI will make things easier

Other industry experts also concurred with Ghose’s views. It is not just the VFX industry, but since generative AI caught the world’s attention with the launch of ChatGPT, concerns about jobs became widespread in many industries, ranging from engineering and programming to many creative fields such as designing.

Nonetheless, experts across various industries share a common perspective that generative AI presently lacks the capacity to fully replace humans. Rather, it functions as a tool to augment human capabilities. It holds true for the VFX industry as well.

Moreover, akin to past technological revolutions, AI is anticipated to reshape the job landscape, opening up new opportunities and avenues for employment.

Notably, even in the VFX industry, the companies have been building proprietary tools for quite some time. Ghose explains that different projects have different requirements, and in some cases, VFX companies develop their own proprietary AI algorithms to support them in that particular project.

Ghose believes that generative AI has the power to improve these existing tools and functions. For instance, “Imagine a landscape featuring a mix of greenery and buildings. When I capture this scene with the camera, visual effects today can enhance and extend the shot. With AI advancements, certain extensions, like modifying the skyline, may become more streamlined.”

GenAI can’t do VFX alone

However, to imagine that an AI model can do the work of a VFX artist or team is far-fetched.

“In the next two years, AI’s capability to serve as a reference or an accelerator, expediting tasks in the realm of imprecise outputs, is expected to enhance productivity significantly. However, achieving the finesse and intricate detailing anticipated in the next one or two years, at least, may still be beyond the reach of AI,” Ghose said.

Other experts whom AIM spoke to on the sidelines of the Bengaluru GAFX 2024 did appear impressed by the quality of the text-to-image AI tools like Midjourney or Stable Diffusion or video-generation tools like Lumiere by Google. However, they, too, believe these tools still do not possess the calibre to support a VFX project independently.

Moreover, another general consensus among experts is that AI currently lacks the logical reasoning and real-world knowledge that artists rely on for creative decision-making and problem-solving.

“When it comes to high-end artistry like what we do, in the short term, the human capability will not be replaced; however, in the long term, nobody knows,” Ghose remarked.

Copyright issues pertain

Ghose also adds that the quality of these tools depends on the data they are trained on. Copyright and AI are still ongoing issues, and many artists in the US and, more famously, The New York Times have sued OpenAI for using their content without permission to train AI models.

Though Ghose believes generative AI will be a game-changer, chances are high that they will develop their own proprietary video generation model by training it with their own rich enterprise data.

“Large organisations like ours, with extensive experience and a wealth of expertise gained from numerous endeavours, will channel this knowledge into proprietary tools tailored to our specific needs. The nature of our work with esteemed clients prohibits us from feeding data into publicly available generative AI tools,” Ghose said.

The post VFX Industry Will Make Their Own ‘Sora-Like’ GenAI Tools appeared first on Analytics India Magazine.

The journey to equality has been a bit of a tedious trek for the LGBTQ+ community in tech (and elsewhere). While some big names in the industry have opened up about their identities, issues like workplace safety and acceptance are real and still hold back many from being themselves.

The evolution within the industry, albeit slow, has been happening with organisations acknowledging the pivotal role diversity, equity, and inclusion (DE&I) play in shaping a more conducive work environment.

Sharing a similar story is Joseleen Princy C, a senior business system analyst at Texas-based marketing and data company Epsilon. Beginning her career as a developer at HCL, she later moved to Cognizant before joining Epsilon where she has been working for almost two years.

A computer science engineer from Chennai, Joseleen identifies as a transwoman and has had her share of troubles with social stigma and workplace biases. But Epsilon provided her with a safe space for coming out and sharing her story.

Internally, she collaborated with the DE&I team to conduct sessions, participated in events, and engaged in media campaigns that reached a broader audience, fostering understanding and acceptance. These efforts collectively contributed to creating a more inclusive and supportive environment for LGBTQ+ individuals within the organisation.

“Epsilon gave me the platform to express myself and advocate for the rights and well-being of my community. I’ve seen a positive impact on the personal and professional growth of my colleagues. Through networking, education, and mentorship, we are empowering our community to thrive in all aspects of their lives,” Princy told AIM in a candid conversation last week.

Epsilon’s D&I Initiatives

According to Princy, the comprehensive support provided by Epsilon during her onboarding, including insurance and medical benefits, has been noteworthy. Of particular significance is the personalised assistance with bank-related documentation. Epsilon took charge of this process, creating a specialised system for efficient compensation processing.

“One thing that I would like to highlight was my bank-related documentation, which Epsilon took over. They generated a special system for me to get my compensation. They also ensured my bank procedures and signatures were in place. This personal attention made me feel valued as an employee and has fostered a sense of trust and transparency within the organisation,” said Princy.

Epsilon aims to create a robust culture of inclusivity, free from biases, to give a respectful and welcoming environment to all team members. The company’s DE&I initiatives are overseen by a dedicated DE&I Council, led by the chief diversity officer, and include Employee Resource Groups (ERGs) that focus on various diversity dimensions.

This commitment is reinforced through mandatory periodic training and workshops covering various topics such as POSH, unconscious bias, ethics, and privacy. “These training ensure that all employees are equipped with the understanding, knowledge, and skills to promote a culture of inclusivity in their daily interactions,” said Princy.

Besides imparting practical knowledge, these initiatives promote empathy and understanding among colleagues, fostering a harmonious and collaborative work environment that reflects the organisation’s dedication to fostering an inclusive workplace.

Beyond training, the company’s efforts encompass recruitment, talent acquisition, learning and development, featuring programs like unconscious bias training. Employee engagement is fostered through surveys and events celebrating diversity. Additionally, Epsilon engages with communities and tracks progress using metrics like workforce representation and participation in DE&I programs.

Addressing Data Gaps for Trans and Non-Binary Individuals

The lack of data on non-binary and transgender individuals stems from challenges in existing gender categorisations. Traditional binary classifications prevalent in data collection often render non-binary and transgender individuals invisible, with estimates suggesting that only 0.1–2% of people identify as transgender or diverge from the cisgender classifications.

These groups face exclusion and discrimination across various life aspects, but limited evidence exists due to the pervasive use of binary gender classifications in data collection instruments.

The minority status of this group, coupled with smaller sample sizes, increases the risk of personal identification in datasets, limiting the ability to derive statistically meaningful insights about non-binary people.

However, the revelation of one’s trans and non-binary identity goes beyond mere employment. There have been several instances where individuals have not been recruited because they chose to reveal their sexual identities.

Beyond employment considerations, the act of coming out as a trans or non-binary person is deeply personal, involving the challenging navigation of societal norms. To address this, Princy suggests support from companies is crucial. This should come in the form of DE&I policies, schemes, and similar supportive systems.

“Improvements should also extend to insurance coverage that includes the third gender, incorporating medical support for necessary procedures. A call for an equal and independent work atmosphere, coupled with opportunities for representation and community welfare awareness, further contributes to fostering inclusivity,” said Princy.

For those struggling to come out, “my advice is to take a leap of faith. Despite initial challenges, there is now widespread awareness and support available. Embracing your true self can lead to a happier, more fulfilling work-life balance,” she said.

Today, companies have open policies, providing a systematic process. She encourages reading policies, and contacting DE&I points of contact, HR, and managers to navigate the systematic process. The industry, she notes, is increasingly supportive, fostering a culture where everyone feels valued and respected, irrespective of background or identity.

The post How Epsilon is Navigating DE&I in Tech appeared first on Analytics India Magazine.

Infosys recently announced its intention to acquire InSemi to strengthen its engineering and R&D capabilities. InSemi, a semiconductor design company brings its expertise in electronic design, platform design, and automation to the table. Last year after multiple misses, Foxconn partnered with HCL Group to build an outsourced assembly and testing unit (OSAT) in India.

L&T, on the other hand, is not seeking partnerships but investing INR 830 crore to set up a wholly owned subsidiary that will be engaged in the business of fabless semiconductor chip design and product ownership.

“Partnering with and acquiring semiconductor firms provides IT companies with the necessary talent and customer access, establishing them as full-stack service providers essential for capturing the growing semiconductor demand,” remarked Pareekh Jain, CEO of EIIR (Engineering IoT R&D) Trend, in an interview with AIM.

Strategic moves in the industry

In the case of Infosys and InSemi, the partnership is expected to significantly bolster Infosys’ engineering R&D capabilities. InSemi is a mammoth in platform design, automation, embedded and software technologies.

InSemi’s expertise will also get Infosys to enhance its chip-to-cloud strategy, providing niche design skills at scale, aligning with Infosys’ heavy investments in AI/automation platforms and industry partnerships.

This acquisition is expected to enable Infosys to offer comprehensive end-to-end product development for its clients, addressing the increasing demand for advanced semiconductor design services integrated with embedded systems, explained Infosys chief Salil Parekh.

TCS also partnered with Renesas Electronics Corporation to open an Innovation Center in Bengaluru and Hyderabad in March 2023. Like Infosys, its focus is to create semiconductor designs and software solutions for various sectors including IoT, infrastructure, industrial, and automotive segments.

They plan to combine their expertise in IoT, manufacturing, telecom, automotive industries and advanced semiconductor designs, and embedded software support to push forward with new semiconductor designs.

Wipro has also been working towards building an alliance with semiconductors for a while. The company acquired Eximius Design in 2020 to improve their VLSI and systems design capabilities. Eximius brought its expertise in semiconductor, software, and systems design to equip Wipro with comprehensive solutions for IoT, Industry 4.0, edge computing, cloud, and 5G applications.

In 2022, Wipro also joined the Intel Foundry Services’ Accelerator Alliance, aiming to expedite the chip design cycle. This partnership was focused on making both complete chips (SoC) and specialised chips (ASIC) for specific uses.

Diversification, from design to manufacturing

HCL Group, on the other hand, is diversifying and foraying into semiconductor manufacturing. In the second attempt by Foxconn, the companies are in talks with Tamil Nadu and Telangana to set up its recently announced semiconductor assembly and testing unit.

“Foxconn and HCL are partnering to establish OSAT operations in India to create a local industry ecosystem of supply chain stability. Foxconn plans to implement its BOL (build-operate-localise) model to benefit local communities through this investment,” a spokesperson for Foxconn said.

The joint venture will be established with $37.2 million investment from Foxconn, securing a 40% stake.

“Indian IT giants have strategically acquired semiconductor engineering firms to bolster their capabilities in this high-growth sector, showing a clear intent to capture the vast potential market for semiconductor outsourcing,” Jain explained.

For HCL Group, the OSAT facility creates infrastructure for chip packaging and testing within India, a crucial step towards self-sufficiency in the semiconductor industry.

Government support and the future

Indian IT companies’ strategic emphasis on semiconductors is also propelled by government initiatives. The Indian government has approved significant financial outlays, including a INR 76,000 crore PLI scheme, to bolster the semiconductor sector and making entry into the segment a lucrative affair.

Larsen & Toubro, a rather new player, made a strategic entry into the semiconductor industry by focusing on fabless chip design rather than manufacturing. R Shankar Raman, L&T’s CFO said at the earnings call last year,

“We are mostly focused on designing automobile and industrial chips. At the moment it will be less investment and low manufacturing, but getting our positioning accessed for the design.”

Each company’s approach differs from the other. While Infosys and Wipro focus on design and development, TCS emphasises innovation through partnerships, HCL on investment in manufacturing capabilities, and L&T on fabless models. The companies in India are working independently in the semiconductor industry, with each one contributing to a different part of the production chain, helping the industry grow as a whole.

In conclusion, the IT sector aims to build semiconductor capabilities in-house, and end-to-end. While L&T are focusing only on chip design, HCL Group is setting up manufacturing facilities. This diversified approach not only aims to attract investment but also seeks to upskill the workforce to meet the industry’s growing demands, as seen through initiatives by HCLTech along with Electronics Sector Skills Council of India (ESSCI)​​​​.

India’s semiconductor sector is set to hit $271.9 billion by 2032, growing at a 25.7% CAGR from 2022. Driven to self sufficiency, Indian IT is pushing hard to fulfil this prediction.

The post Why is Indian IT Investing in Semiconductors appeared first on Analytics India Magazine.

Armilla wants to give companies a warranty for AI Kyle Wiggers 23 hours

There’s a lot that can go wrong with GenAI — especially third-party GenAI. It makes stuff up. It’s biased and toxic. And it can run afoul of copyright rules. According to a recent survey from MIT Sloan Management Review and Boston Consulting Group, third-party AI tools are responsible for over 55% of AI-related failures in organizations.

So it’s not surprising, exactly, that some companies are wary of adopting the tech just yet.

But what if GenAI came with a warranty?

That’s the business idea Karthik Ramakrishnan, an entrepreneur and electrical engineer, came up with several years ago while working at Deloitte as a senior manager. He’d co-founded two “AI-first” companies, Gallop Labs and Blu Trumpet, and eventually came to realize that trust — and being able to quantify risk — was holding back the adoption of AI.

“Right now, nearly every enterprise is looking for ways to implement AI to increase efficiency and keep up with the market,” Ramakrishnan told TechCrunch in an email interview. “To do this, many are turning to third-party vendors and implementing their AI models without a complete understanding of the quality of the products … AI is advancing at such a rapid pace that the risks and harms are always evolving.”

So Ramakrishnan teamed up with Dan Adamson, an expert in search algorithms and two-time startup founder, to start Armilla AI, which provides warranties on AI models to corporate customers.

How, you might be wondering, can Armilla do this given that most models are black boxes or otherwise gated behind licenses, subscriptions and APIs? I had the same question. Through benchmarking, was Ramakrishnan’s answer — and a careful approach to customer acquisition.

Armilla takes a model — whether open source or proprietary — and conducts assessments to “verify its quality,” informed by the global AI regulatory landscape. The company tests for things like hallucinations, racial and gender bias and fairness, general robustness and security across an array of theoretical applications and use cases, leveraging its in-house assessment technology.

Image Credits: Armilla

If the model passes muster, Armilla backs it with its warranty, which reimburses the model’s buyer for any fee they paid to use the model.

“What we really offer enterprises is confidence in the technology they’re procuring from third-party AI vendors,” Ramakrishnan said. “Enterprises can come to us and have us run assessments on the vendors they’re looking to use. Just like the penetration testing they would do for new technology, we perform penetration testing for AI.”

I asked Ramakrishnan, by the way, whether there were any models Armilla wouldn’t test for ethical reasons — say a facial recognition algorithm from a vendor known to do business with questionable actors. He said:

“It would not only be against our ethics, but against our business model, which is predicated on trust, to produce assessments and reports that provide false confidence in AI models that are problematic for a client and society. From a legal standpoint, we aren’t going to take on clients for models that are prohibited by the EU, that have been banned, which is the case for some facial recognition and biometric categorization systems, for example — but applications that fall into the ‘higher risk’ category as defined by the EU AI Act, yes.”

Now, the concept of warranties and policy coverage for AI isn’t new — a fact that surprised this writer, frankly. Last year, Munich Re debuted an insurance product, aiSure, designed to protect against losses from potentially unreliable AI models by running the models through benchmarks similar to Armilla’s. Outside of warranties, a growing number of vendors, including OpenAI, Microsoft and AWS, offer protections pertaining to copyright violations that might arise from deployment of their AI tools.

But Ramakrishnan claims that Armilla’s approach is unique.

“Our assessment touches on a wide range of areas, including KPIs, processes, performance, data quality, and qualitative and quantitative criteria, and we do it at a fraction of the cost and time,” he added. “We assess AI models based on requirements set out in legislation such as EU AI Act or the AI hiring bias law in NYC — NYC Local Law 144 — and other state regulations, such as Colorado’s proposed AI quantitative testing regulation or New York’s insurance circular on the use of AI in underwriting or pricing. We’re also ready to conduct assessments required by other emerging regulations as they come into play, such as Canada’s AI and Data Act.”

Armilla, which launched coverage in late 2023, backed by carriers Swiss Re, Greenlight Re and Chaucer, claims to have 10 or so customers, including a healthcare company applying GenAI to process medical records. Ramakrishnan tells me that Armilla’s client base has been growing 2x month over month since Q4 2023.

“We’re serving two main audiences: enterprises and third-party AI vendors,” Ramakrishnan said. “Enterprises use our warranty to establish protection for the third-party AI vendors they’re procuring. Third-party vendors use our warranty as a stamp of approval that their product is trustworthy, which helps to shorten their sales cycles.”

Warranties for AI make intuitive sense. But part of me wonders whether Armilla will be able to keep up with fast-shifting AI policy (e.g. New York City’s hiring algorithm bias law, the EU AI Act, etc.), which could put it on the hook for sizeable payouts if its assessments — and contracts — aren’t bulletproof.

Ramakrishnan brushed aside this concern.

“Regulation is rapidly developing in many jurisdictions independently,” he said, “and it’ll be critical to understand the nuances of legislation around the world. There’s no ‘one-size-fits-all’ that we can apply as a global standard, so we need to stitch it all together. This is challenging — but has the benefit of creating a ‘moat’ for us.”

Armilla — based in Toronto, with 13 employees — recently raised $4.5 million in a seed round led by Mistral (not to be confused with the AI startup of the same name) with participation from Greycroft, Differential Venture Capital, Mozilla Ventures, Betaworks Ventures, MS&AD Ventures, 630 Ventures, Morgan Creek Digital, Y Combinator, Greenlight Re and Chaucer. Bringing its total raised to $7 million, Ramakrishnan said that the proceeds will be put toward expanding Armilla’s existing warranty offering as well as introducing new products.

“Insurance will play the biggest role in addressing AI risk, and Armilla is at the forefront of developing insurance products that will allow companies to deploy AI solutions safely,” Ramakrishnan said.

OpenAI chief Sam Altman brought Cred founder Kunal Shah’s imaginative prompt to life by creating a video featuring a bicycle race on the ocean using Sora.

Altman had a busy Thursday as OpenAI introduced its text-to-video generation model, Sora. Eager to engage with users, Altman took to X, urging them to contribute prompt ideas. These user-provided prompts served as the creative fuel for Altman, inspiring the videos he crafted.

Shah’s prompt read, “A bicycle race on the ocean with different animals as athletes riding the bicycles with a drone camera view.”

OpenAI’s Sora is designed to understand and simulate complex scenes, featuring multiple characters, specific motions, and intricate details of the subject and background. The model not only interprets user prompts accurately but also ensures the persistence of characters and visual style throughout the generated video.

One of Sora’s standout features is its ability to take existing still images and breathe life into them, animating the content with precision and attention to detail. Additionally, it can extend or fill in missing frames in an existing video, showcasing its versatility in manipulating visual data.

Sora builds on past research in DALL·E and GPT models. It uses the recaptioning technique from DALL·E 3, which involves generating highly descriptive captions for the visual training data.

While Sora’s capabilities are impressive, OpenAI acknowledges certain weaknesses, such as challenges in accurately simulating the physics of complex scenes and occasional confusion regarding spatial details in prompts.

The post Sam Altman Brings CRED Founder Kunal Shah’s Wild Imagination to Life with Sora appeared first on Analytics India Magazine.

The global market for AI video generators, valued at $472.9 million in 2022, is projected to grow at 19.7% from 2023 to 2030. The escalating demand for video content, driven by the increasing popularity of visual environments in businesses, has resulted in approximately 80% of online traffic being attributed to videos.

In the past one-and-a-half years, we have seen several startups tapping into the space with interesting models involving video creation from diverse sources like text, PowerPoint, or spreadsheets.

London-based Synthesia was one of the early companies to offer an AI-powered platform featuring Avatars, supporting over 60 languages, a screen recorder, and templates. However, Synthesia is not the only video generation platform to have gained widespread adoption.

Let’s take a look at some of the Sythesia AI alternatives.

Best Systhesia AI Video Generator Alternatives

• Speechify
• Runway
• HeyGen
• DeepBrain AI
• VEED
• Colossyan

Speechify

Launched in 2020 by Cliff Weitzman, the Speechify AI video generator is designed for marketers, educators, and content creators. This tool uses AI to transform written scripts into compelling videos, eliminating the need for complex software or editing skills.

You need to input your text, select a video style (such as slideshow or whiteboard), and let Speechify’s AI handle the rest. It generates personalised visuals, incorporates AI-powered voiceovers with diverse accents, and even includes royalty-free music.

Pricing options range from a free plan with limited features to paid plans starting at $27 per month, providing longer videos, more voiceover options, and access to a vast library of stock footage. A free trial is also available for experimentation.

While Speechify excels in creating basic explainer videos and social media content, it may not be suitable for intricate narratives or highly customised projects. Users have reported occasional AI voice glitches and limitations in video style choices.

Runway

Applied AI research company Runway’s flagship tool, Gen-2, enables users to create impressive, original videos from simple text prompts. Its capabilities extend beyond text so that users can manipulate images or existing videos with over 30 AI tools through a user-friendly drag-and-drop interface.

The pricing structure accommodates both hobbyists and professionals, with a free-tier offering basic features and paid plans providing more robust tools and higher resolutions. With its interactive interface, a number of features, and commitment to responsible AI development, Runway aims to empower anyone to become a video storyteller.

HeyGen

Like other AI video generators, HeyGen also allows users to convert scripts into engaging videos but with AI-powered avatars, eliminating the need for cameras and crews. The platform offers a diverse library of customisable AI avatars, text overlays, and background images, catering to various purposes like explainer videos, corporate training, e-commerce ads, and social media content.

Users can also upload their own graphics and audio for a personalised touch. HeyGen’s pricing starts with a free plan but paid plans, such as the Pro plan at $49/month, provide access to premium avatars, extended video lengths, and additional editing options. Despite its limitations, HeyGen has gained popularity for its user-friendly interface and cost-effectiveness, attracting both businesses and individuals.

DeepBrain AI

Established in 2016 by Eric Jang, DeepBrain AI focuses on streamlining video creation through AI. The standout product, AI Studios, is an online platform enabling users to swiftly generate polished videos from text scripts. Noteworthy features include a selection of over 100 realistic AI avatars, multilingual natural speech, text-to-video conversion, content transformation from various sources, and easy editing through a drag-and-drop interface.

AI avatars use advanced text-to-speech technology to speak various languages, enabling the generation of videos for global audiences in languages such as English, Spanish, Chinese, Korean, and more. It also offers significant time and cost savings, and is up to 80% more cost-effective and faster compared to traditional video production methods.

DeepBrain AI offers diverse pricing plans, ranging from a free trial to enterprise-level subscriptions. The platform has garnered attention by democratising video creation, forming partnerships with notable companies like Microsoft, Lenovo and NVIDIA.

VEED

UK-based SaaS startup VEED.IO’s overarching mission is to democratise video creation by offering tools that are both simple and powerful. Some of the key AI powered features include a script generator, text-to-speech functionality with realistic narrators in multiple languages, an image generator to translate scripts into visuals, and a stock media library providing access to a vast collection of royalty-free video clips, images, and music tracks.

The free plan offers a taste of the platform’s capabilities with certain export limitations, while paid plans, starting at $12 per month.

Colossyan

Colossyan, started in 2020 by Dominik Kovacs, stands out as a Hungarian startup specialising in AI-driven video creation for workplace learning and development. The platform caters to organisations and educators seeking cost-effective, engaging, and multilingual AI-generated video content for training purposes.

Its main product, Colossyan Creator, offers features like text-to-video, multiple avatars for interactive scenarios, automatic translation to over 70 languages, and tools for branding and collaboration. Colossyan Creator facilitates the rapid creation of engaging training materials, product demos, and internal communications, aligning with their mission to make video creation accessible and efficient.

Pricing options include free, Starter, Pro, and Enterprise plans, with custom pricing available for high-volume users.

Popular Posts

• ChatGPT Alternatives
• Midjourney Alternatives
• GitHub Copilot Alternatives
• Google Generative AI Alternatives
• LangChain Alternatives for Building AI Agents

The post Top 6 Synthesia AI Alternative in 2024 appeared first on Analytics India Magazine.

Google’s new Gemini model can analyze an hour-long video — but few people can use it Kyle Wiggers 20 hours

Last October, a research paper published by a Google data scientist, the CTO of Databricks Matei Zaharia and UC Berkeley professor Pieter Abbeel posited a way to allow GenAI models — i.e. models along the lines of OpenAI’s GPT-4 and ChatGPT — to ingest far more data than was previously possible. In the study, the co-authors demonstrated that, by removing a major memory bottleneck for AI models, they could enable models to process millions of words as opposed to hundreds of thousands — the maximum of the most capable models at the time.

AI research moves fast, it seems.

Today, Google announced the release of Gemini 1.5 Pro, the newest member of its Gemini family of GenAI models. Designed to be a drop-in replacement for Gemini 1.0 Pro (which formerly went by “Gemini Pro 1.0” for reasons known only to Google’s labyrinthine marketing arm), Gemini 1.5 Pro is improved in a number of areas compared with its predecessor, perhaps most significantly in the amount of data that it can process.

Gemini 1.5 Pro can take in ~700,000 words, or ~30,000 lines of code — 35x the amount Gemini 1.0 Pro can handle. And — the model being multimodal — it’s not limited to text. Gemini 1.5 Pro can ingest up to 11 hours of audio or an hour of video in a variety of different languages.

Image Credits: Google

To be clear, that’s an upper bound.

The version of Gemini 1.5 Pro available to most developers and customers starting today (in a limited preview) can only process ~100,000 words at once. Google’s characterizing the large-data-input Gemini 1.5 Pro as “experimental,” allowing only developers approved as part of a private preview to pilot it via the company’s GenAI dev tool AI Studio. Several customers using Google’s Vertex AI platform also have access to the large-data-input Gemini 1.5 Pro — but not all.

Still, VP of research at Google DeepMind Oriol Vinyals heralded it as an achievement.

“When you interact with [GenAI] models, the information you’re inputting and outputting becomes the context, and the longer and more complex your questions and interactions are, the longer the context the model needs to be able to deal with gets,” Vinyals said during a press briefing. “We’ve unlocked long context in a pretty massive way.”

Big context

A model’s context, or context window, refers to input data (e.g. text) that the model considers before generating output (e.g. additional text). A simple question — “Who won the 2020 U.S. presidential election?” — can serve as context, as can a movie script, email or e-book.

Models with small context windows tend to “forget” the content of even very recent conversations, leading them to veer off topic — often in problematic ways. This isn’t necessarily so with models with large contexts. As an added upside, large-context models can better grasp the narrative flow of data they take in and generate more contextually rich responses — hypothetically, at least.

There have been other attempts at — and experiments on — models with atypically large context windows.

AI startup Magic claimed last summer to have developed a large language model (LLM) with a 5 million-token context window. Two papers in the past year detail model architectures ostensibly capable of scaling to a million tokens — and beyond. (“Tokens” are subdivided bits of raw data, like the syllables “fan,” “tas” and “tic” in the word “fantastic.”) And recently, a group of scientists hailing from Meta, MIT and Carnegie Mellon developed a technique that they say removes the constraint on model context window size altogether.

But Google is the first to make a model with a context window of this size commercially available, beating the previous leader Anthropic’s 200,000-token context window — if a private preview counts as commercially available.

Image Credits: Google

Gemini 1.5 Pro’s maximum context window is 1 million tokens, and the version of the model more widely available has a 128,000-token context window, the same as OpenAI’s GPT-4 Turbo.

So what can one accomplish with a 1 million-token context window? Lots of things, Google promises — like analyzing a whole code library, “reasoning across” lengthy documents like contracts, holding long conversations with a chatbot and analyzing and comparing content in videos.

During the briefing, Google showed two prerecorded demos of Gemini 1.5 Pro with the 1 million-token context window enabled.

In the first, the demonstrator asked Gemini 1.5 Pro to search the transcript of the Apollo 11 moon landing telecast — which comes to around 402 pages — for quotes containing jokes, and then to find a scene in the telecast that looked similar to a pencil sketch. In the second, the demonstrator told the model to search for scenes in “Sherlock Jr.,” the Buster Keaton film, going by descriptions and another sketch.

Image Credits: Google

Gemini 1.5 Pro successfully completed all the tasks asked of it, but not particularly quickly. Each took between ~20 seconds and a minute to process — far longer than, say, the average ChatGPT query.

Image Credits: Google

Vinyals says that the latency will improve as the model’s optimized. Already, the company’s testing a version of Gemini 1.5 Pro with a 10 million-token context window.

“The latency aspect [is something] we’re … working to optimize — this is still in an experimental stage, in a research stage,” he said. “So these issues I would say are present like with any other model.”

Me, I’m not so sure latency that poor will be attractive to many folks — much less paying customers. Having to wait minutes at a time to search across a video doesn’t sound pleasant — or very scalable in the near term. And I’m concerned how the latency manifests in other applications, like chatbot conversations and analyzing codebases. Vinyals didn’t say — which doesn’t instill much confidence.

My more optimistic colleague Frederic Lardinois pointed out that the overall time savings might just make the thumb twiddling worth it. But I think it’ll depend very much on the use case. For picking out a show’s plot points? Perhaps not. But for finding the right screengrab from a movie scene you only hazily recall? Maybe.

Other improvements

Beyond the expanded context window, Gemini 1.5 Pro brings other, quality-of-life upgrades to the table.

Google’s claiming that — in terms of quality — Gemini 1.5 Pro is “comparable” to the current version of Gemini Ultra, Google’s flagship GenAI model, thanks to a new architecture comprised of smaller, specialized “expert” models. Gemini 1.5 Pro essentially breaks down tasks into multiple subtasks and then delegates them to the appropriate expert models, deciding which task to delegate based on its own predictions.

MoE isn’t novel — it’s been around in some form for years. But its efficiency and flexibility has made it an increasingly popular choice among model vendors (see: the model powering Microsoft’s language translation services).

Now, “comparable quality” is a bit of a nebulous descriptor. Quality where it concerns GenAI models, especially multimodal ones, is hard to quantify — doubly so when the models are gated behind private previews that exclude the press. For what it’s worth, Google claims that Gemini 1.5 Pro performs at a “broadly similar level” compared to Ultra on the benchmarks the company uses to develop LLMs while outperforming Gemini 1.0 Pro on 87% of those benchmarks. (I’ll note that outperforming Gemini 1.0 Pro is a low bar.)

Pricing is a big question mark.

During the private preview, Gemini 1.5 Pro with the 1 million-token context window will be free to use, Google says. But the company plans to introduce pricing tiers in the near future that start at the standard 128,000 context window and scale up to 1 million tokens.

I have to imagine the larger context window won’t come cheap — and Google didn’t allay fears by opting not to reveal pricing during the briefing. If pricing’s in line with Anthropic’s, it could cost $8 per million prompt tokens and $24 per million generated tokens. But perhaps it’ll be lower; stranger things have happened! We’ll have to wait and see.

I wonder, too, about the implications for the rest of the models in the Gemini family, chiefly Gemini Ultra. Can we expect Ultra model upgrades roughly aligned with Pro upgrades? Or will there always be — as there is now — an awkward period where the available Pro models are superior performance-wise to the Ultra models, which Google’s still marketing as the top of the line in its Gemini portfolio?

Chalk it up to teething issues if you’re feeling charitable. If you’re not, call it like it is: darn confusing.

Whatfix, the global leader among digital adoption platforms (DAP), today announced the launch of yet another new product called Mirror, which is set to revolutionise systems training and product showcasing.

Mirror creates hyper-realistic and interactive replicas of web applications for immersive training and product demonstrations without any of the risks of live system engagement, is slated for Beta release in Q2’24.

IT departments will cut down significant infrastructure and manpower costs associated with maintaining additional application environments. Several large enterprises, including Fortune 500 companies, have realized value during the initial trials of Mirror, the company said.

“Whatfix remains dedicated to improving user experiences, and with the launch of Mirror, we not only added a cutting-edge product line to our portfolio but also strengthened our position in Digital Adoption Platforms (DAP) and Analytics. This expansion underscores our commitment to revolutionizing user experiences, reducing costs, and accelerating the return on investment for digital transformations”, said Khadim Batti, Whatfix CEO and co-founder.

Whatfix registered a second year of top decile Year-over-Year (YoY) 45% growth in Annual Recurring Revenue (ARR) and a substantial 35% YoY increase in new revenue generated from existing customers.

The company also celebrated the successful closure of six deals exceeding USD 1 million, underscoring the need for organization-wide DAP implementations in large enterprises. A 40% surge in the Average Revenue Per Account, affirmed the increasing trust of existing customers.

The post Whatfix Launches Immersive Training Tool Mirror appeared first on Analytics India Magazine.