DeepSeek’s R1 and OpenAI’s Deep Research just redefined AI — RAG, distillation, and custom models will never be the same - Related to how, distillation,, custom, accessible, bi:

DeepSeek’s R1 and OpenAI’s Deep Research just redefined AI — RAG, distillation, and custom models will never be the same

Things are moving quickly in AI — and if you’re not keeping up, you’re falling behind.

Two recent developments are reshaping the landscape for developers and enterprises alike: DeepSeek’s R1 model release and OpenAI’s new Deep Research product. Together, they’re redefining the cost and accessibility of powerful reasoning models, which has been well reported on. Less talked about, however, is how they’ll push companies to use techniques like distillation, supervised fine-tuning (SFT), reinforcement learning (RL) and retrieval-augmented generation (RAG) to build smarter, more specialized AI applications.

After the initial excitement around the amazing achievements of DeepSeek begins to settle, developers and enterprise decision-makers need to consider what it means for them. From pricing and performance to hallucination risks and the importance of clean data, here’s what these breakthroughs mean for anyone building AI today.

Cheaper, transparent, industry-leading reasoning models – but through distillation.

The headline with DeepSeek-R1 is simple: It delivers an industry-leading reasoning model at a fraction of the cost of OpenAI’s o1. Specifically, it’s about 30 times cheaper to run, and unlike many closed models, DeepSeek offers full transparency around its reasoning steps. For developers, this means you can now build highly customized AI models without breaking the bank — whether through distillation, fine-tuning or simple RAG implementations.

Distillation, in particular, is emerging as a powerful tool. By using DeepSeek-R1 as a “teacher model,” companies can create smaller, task-specific models that inherit R1’s superior reasoning capabilities. These smaller models, in fact, are the future for most enterprise companies. The full R1 reasoning model can be too much for what companies need — thinking too much, and not taking the decisive action companies need for their specific domain applications.

“One of the things that no one is really talking about, certainly in the mainstream media, is that, actually, reasoning models are not working that well for things like agents,” stated Sam Witteveen, a machine learning (ML) developer who works on AI agents that are increasingly orchestrating enterprise applications.

As part of its release, DeepSeek distilled its own reasoning capabilities onto a number of smaller models, including open-source models from Meta’s Llama family and Alibaba’s Qwen family, as described in its paper. It’s these smaller models that can then be optimized for specific tasks. This trend toward smaller, fast models to serve custom-built needs will accelerate: Eventually there will be armies of them.

“We are starting to move into a world now where people are using multiple models. They’re not just using one model all the time,” expressed Witteveen. And this includes the low-cost, smaller closed-sourced models from Google and OpenAI as well. “The means that models like Gemini Flash, GPT-4o Mini, and these really cheap models actually work really well for 80% of use cases.”.

If you work in an obscure domain, and have resources: Use SFT….

After the distilling step, enterprise companies have a few options to make sure the model is ready for their specific application. If you’re a corporation in a very specific domain, where details are not on the web or in books — which large language models (LLMs) typically train on — you can inject it with your own domain-specific data sets, with SFT. One example would be the ship container-building industry, where specifications, protocols and regulations are not widely available.

DeepSeek showed that you can do this well with “thousands” of question-answer data sets. For an example of how others can put this into practice, IBM engineer Chris Hay demonstrated how he fine-tuned a small model using his own math-specific datasets to achieve lightning-fast responses — outperforming OpenAI’s o1 on the same tasks (View the hands-on video here.).

Additionally, companies wanting to train a model with additional alignment to specific preferences — for example, making a customer support chatbot sound empathetic while being concise — will want to do some RL. This is also good if a organization wants its chatbot to adapt its tone and recommendation based on user feedback. As every model gets good at everything, “personality” is going to be increasingly big, Wharton AI professor Ethan Mollick stated on X.

These SFT and RL steps can be tricky for companies to implement well, however. Feed the model with data from one specific domain area, or tune it to act a certain way, and it suddenly becomes useless for doing tasks outside of that domain or style.

For most companies, RAG will be good enough.

For most companies, however, RAG is the easiest and safest path forward. RAG is a relatively straight-forward process that allows organizations to ground their models with proprietary data contained in their own databases — ensuring outputs are accurate and domain-specific. Here, an LLM feeds a user’s prompt into vector and graph databases to search information relevant to that prompt. RAG processes have gotten very good at finding only the most relevant content.

This approach also helps counteract some of the hallucination issues associated with DeepSeek, which currently hallucinates 14% of the time compared to 8% for OpenAI’s o3 model, , a vendor that helps companies with the RAG process.

This distillation of models plus RAG is where the magic will come for most companies. It has become so incredibly easy to do, even for those with limited data science or coding expertise. I personally downloaded the DeepSeek distilled [website] Qwen model, the smallest one, so that it could fit nicely on my Macbook Air. I then loaded up some PDFs of job applicant resumes into a vector database, then asked the model to look over the applicants to tell me which ones were qualified to work at VentureBeat. (In all, this took me 74 lines of code, which I basically borrowed from others doing the same).

I loved that the Deepseek distilled model showed its thinking process behind why or why not it recommended each applicant — a transparency that I wouldn’t have gotten easily before Deepseek’s release.

In my recent video discussion on DeepSeek and RAG, I walked through how simple it has become to implement RAG in practical applications, even for non-experts. Witteveen also contributed to the discussion by breaking down how RAG pipelines work and why enterprises are increasingly relying on them instead of fully fine-tuning models. (Watch it here).

OpenAI Deep Research: Extending RAG’s capabilities — but with caveats.

While DeepSeek is making reasoning models cheaper and more transparent, OpenAI’s Deep Research represents a different but complementary shift. It can take RAG to a new level by crawling the web to create highly customized research. The output of this research can then be inserted as input into the RAG documents companies can use, alongside their own data.

This functionality, often referred to as agentic RAG, allows AI systems to autonomously seek out the best context from across the internet, bringing a new dimension to knowledge retrieval and grounding.

Open AI’s Deep Research is similar to tools like Google’s Deep Research, Perplexity and [website], but OpenAI tried to differentiate its offering by suggesting its superior chain-of-thought reasoning makes it more accurate. This is how these tools work: A organization researcher requests the LLM to find all the information available about a topic in a well-researched and cited analysis. The LLM then responds by asking the researcher to answer another 20 sub-questions to confirm what is wanted. The research LLM then goes out and performs 10 or 20 web searches to get the most relevant data to answer all those sub-questions, then extract the knowledge and present it in a useful way.

However, this innovation isn’t without its challenges. Vectara CEO Amr Awadallah cautioned about the risks of relying too heavily on outputs from models like Deep Research. He questions whether indeed it is more accurate: “It’s not clear that this is true,” Awadallah noted. “We’re seeing articles and posts in various forums saying no, they’re getting lots of hallucinations still, and Deep Research is only about as good as other solutions out there on the market.”.

In other words, while Deep Research offers promising capabilities, enterprises need to tread carefully when integrating its outputs into their knowledge bases. The grounding knowledge for a model should come from verified, human-approved reports to avoid cascading errors, Awadallah noted.

The cost curve is crashing: Why this matters.

The most immediate impact of DeepSeek’s release is its aggressive price reduction. The tech industry expected costs to come down over time, but few anticipated just how quickly it would happen. DeepSeek has proven that powerful, open models can be both affordable and efficient, creating opportunities for widespread experimentation and cost-effective deployment.

Awadallah emphasized this point, noting that the real game-changer isn’t just the training cost — it’s the inference cost, which for DeepSeek is about 1/30th of OpenAI’s o1 or o3 for inference cost per token. “The margins that OpenAI, Anthropic and Google Gemini were able to capture will now have to be squished by at least 90% because they can’t stay competitive with such high pricing,” mentioned Awadallah.

Not only that, those costs will continue to go down. Anthropic CEO Dario Amodei expressed lately that the cost of developing models continues to drop at around a 4x rate each year. It follows that the rate that LLM providers charge to use them will continue to drop as well.

“I fully expect the cost to go to zero,” noted Ashok Srivastava, CDO of Intuit, a firm that has been driving AI hard in its tax and accounting software offerings like TurboTax and Quickbooks. “…and the latency to go to zero. They’re just going to be commodity capabilities that we will be able to use.”.

This cost reduction isn’t just a win for developers and enterprise people; it’s a signal that AI innovation is no longer confined to big labs with billion-dollar budgets. The barriers to entry have dropped, and that’s inspiring smaller companies and individual developers to experiment in ways that were previously unthinkable. Most importantly, the models are so accessible that any business professional will be using them, not just AI experts, stated Srivastava.

DeepSeek’s disruption: Challenging “Big AI’s” stronghold on model development.

Most importantly, DeepSeek has shattered the myth that only major AI labs can innovate. For years, companies like OpenAI and Google positioned themselves as the gatekeepers of advanced AI, spreading the belief that only top-tier PhDs with vast resources could build competitive models.

DeepSeek has flipped that narrative. By making reasoning models open and affordable, it has empowered a new wave of developers and enterprise companies to experiment and innovate without needing billions in funding. This democratization is particularly significant in the post-training stages — like RL and fine-tuning — where the most exciting developments are happening.

DeepSeek exposed a fallacy that had emerged in AI — that only the big AI labs and companies could really innovate. This fallacy had forced a lot of other AI builders to the sidelines. DeepSeek has put a stop to that. It has given everyone inspiration that there’s a ton of ways to innovate in this area.

The Data imperative: Why clean, curated data is the next action-item for enterprise companies.

While DeepSeek and Deep Research offer powerful tools, their effectiveness ultimately hinges on one critical factor: Data quality. Getting your data in order has been a big theme for years, and has accelerated over the past nine years of the AI era. But it has become even more key with generative AI, and now with DeepSeek’s disruption, it’s absolutely key.

Hilary Packer, CTO of American Express, underscored this in an interview with VentureBeat: “The aha! moment for us, honestly, was the data. You can make the best model selection in the world… but the data is key. Validation and accuracy are the holy grail right now of generative AI.”.

This is where enterprises must focus their efforts. While it’s tempting to chase the latest models and techniques, the foundation of any successful AI application is clean, well-structured data. Whether you’re using RAG, SFT or RL, the quality of your data will determine the accuracy and reliability of your models.

And, while many companies aspire to perfect their entire data ecosystems, the reality is that perfection is elusive. Instead, businesses should focus on cleaning and curating the most critical portions of their data to enable point AI applications that deliver immediate value.

Related to this, a lot of questions linger around the exact data that DeepSeek used to train its models on, and this in turn raises questions about the inherent bias of the knowledge stored in its model weights. But that’s no different from questions around other open-source models, such as Meta’s Llama model series. Most enterprise individuals have found ways to fine-tune or ground the models with RAG enough so that they can mitigate any problems around such biases. And that’s been enough to create serious momentum within enterprise companies toward accepting open source, indeed even leading with open source.

Similarly, there’s no question that many companies will be using DeepSeek models, regardless of the fear around the fact that the corporation is from China. Although it’s also true that a lot of companies in highly regulated industries such as finance or healthcare are going to be cautious about using any DeepSeek model in any application that interfaces directly with people, at least in the short-term.

Conclusion: The future of enterprise AI Is open, affordable and data-driven.

DeepSeek and OpenAI’s Deep Research are more than just new tools in the AI arsenal — they’re signals of a profound shift where enterprises will be rolling out masses of purpose-built models, extremely affordably, competent and grounded in the firm’s own data and approach.

For enterprises, the message is clear: The tools to build powerful, domain-specific AI applications are at your fingertips. You risk falling behind if you don’t leverage these tools. But real success will come from how you curate your data, leverage techniques like RAG and distillation and innovate beyond the pre-training phase.

As AmEx’s Packer put it: The companies that get their data right will be the ones leading the next wave of AI innovation.

OpenAI CEO Sam Altman is set to visit India for the second time this Wednesday. His Asia tour has already seen several product launches so far and his...

Ever since AI/ML workloads came into play, companies have been exploring new data container orchestration platforms and slowly moving away from Kubern...

Day one of MLDS 2025, India’s biggest GenAI summit for developers, hosted by AIM Media House, was a day filled with energy, excitement, and forward-th...

How Self Evolving Agents Pose Risks for the Future Workforce

Are you ready for a world where AI agents work alongside humans, not just as tools but as decision-makers? Imagine walking into your favourite coffee shop and overhearing a conversation: “I’ve built an agent.”.

This is what Rahul Bhattacharya, AI leader, GDS Consulting, EY, spoke about at MLDS 2025, discussing the role of the self evolving agentic workforce of the future, and how assessing the risk is equally critical as to measuring the benefits.

Bhattacharya explained that for a system to be considered an agent, it must have certain abilities. It should be able to interact with its environment by observing what’s happening around it and taking actions. It must also understand changes in the world, recognising what happens after it makes a move.

“A key ability is making decisions, where the agent chooses the best action based on set rules, goals, or rewards. Over time, it should learn from past experiences and feedback to improve its performance,” Bhattacharya stated.

Additionally, an agent must balance new ideas with proven methods, exploring different approaches while still using what works best.

Giving an example of self-driving cars, which senses their surroundings, follow traffic rules, make decisions, and “learns from real-time data,” Bhattacharya pointed out that an agent also needs to have agency, which is the ability to make choices and not just follow a fixed path, which is the risk factor since they become unpredictable.

He mentioned that one major difference of current AI agents with what was discussed with LLMs a year back is “Tools vs. Actions.”.

A tool has a fixed, predictable output, like a calculator, while an action is more flexible and can lead to different results, such as an AI assistant making a complex decision.

Another key aspect is planning and memory as AI agents can break tasks into smaller steps (sub-goal decomposition) and use memory, both short-term (within a task) and long-term (learning over time).

As Bhattacharya observed that the workforce of the future will not just be made up of humans but will also include “teams of AI agents working alongside people.” Instead of hiring only humans, companies will begin to deploy AI agents for tasks.

Some of these tasks will go to deterministic tools that follow fixed processes, while others will be handled by AI agents that can make flexible decisions. Just like humans, these agents will need knowledge—both general skills and organization-specific information about internal processes.

This shift is also creating new job roles. “Knowledge Harvesters” will be responsible for collecting and documenting human knowledge so AI agents can use it, while “Flow Engineers” will decide which tasks should be assigned to AI agents, which should remain as tools, and how everything should work together.

This brought Bhattacharya to talk about AGI. He mentioned that instead of a single, super-intelligent AI, there could be a “network of self-evolving AI agents” that can “self-spawn” (create new agents) and “self-train” (learn new skills).

He described a future where an AI system starts with no agents, but as tasks arise, it creates a new agent to handle them, leading to continuous growth and learning—possibly even true AGI.

However, this progress also comes with risks. AI must have “agency”, meaning the ability to make decisions, but “agency creates risk because it is not deterministic… It might take actions that do not align with our morals, ethics, or business policies.”To keep AI under control, observability is crucial. Just like how airplanes rely on autopilot but still require human pilots for safety, AI systems need oversight to ensure they make the right choices within safe boundaries.

Sauce Labs Inc, a leading provider of software testing solutions, has appointed Prince Kohli as its new CEO. Kohli, an expert in AI-driven solutions a...

Developer platform GitHub has introduced Agent Mode for GitHub Copilot, giving its AI-powered coding assistant the ability to iterate on its own code,...

Language models (LMs) trained to predict the next word given input text are the key technology for many applications [1, 2]. In Gboard, LMs are used t...

When AI Met BI: How Amazon QuickSight is Making Data More Accessible

Amazon Q Business, first introduced at the Amazon Web Services (AWS) re:Invent in 2023, has evolved over the past 12 months to become a comprehensive AI assistant that can answer questions, summarise content, generate visuals, and automate tasks – all based on an organisation’s data.

QuickSight is Amazon’s completely managed, cloud-native business intelligence (BI) service that is revolutionising the way an organisation interacts with their data and empowers teams at all levels to adopt a data-driven culture.

What if your business could unlock the full potential of its data, transforming it into insights that drive smarter decisions at the speed of thought? That’s exactly what Amazon QuickSight delivers. Its functions include machine learning–powered insights, natural language queries through QuickSight Q, and predictive analytics that can democratise data and allow customers, whether technical or non-technical, to access and act on the data.

Tracy Daugherty, general manager of Amazon QuickSight, has been the driving force behind this evolution and has guided the platform through modern data analytics challenges. QuickSight has never been improved for scale. With support for multi-tenant deployments for enterprise needs, the offering now easily integrates with other AWS services, including S3, Redshift, and Athena.

At AWS re:Invent 2024, AIM interviewed Daugherty and had a detailed discussion about the platform’s journey and its vision for the future.

Reflecting on the platform’s early days, Tracy unveiled, “When I joined seven years ago, Amazon QuickSight was architecture-rich but feature-poor.” Despite its potential cloud-native architecture, Amazon QuickSight faced stiff competition from established vendors such as Microsoft Power BI and Tableau.

Tracy and his colleagues identified an opportunity to differentiate Amazon QuickSight by emphasising accessibility. “Amazon QuickSight was built to be more than a dashboard tool,” he explained. “It’s about empowering everyone in the organisation with insights, whether you’re a business executive, developer, or frontline worker.”.

This vision transformed Amazon QuickSight from a basic reporting tool to a full-fledged, self-service BI platform that could support use cases ranging from dashboards and reporting to embedded analytics. “Our purpose was always to democratise data access,” Tracy mentioned. “We wanted to build a tool that would suit everybody – from technical analysts to those who have no technical background.”.

One of QuickSight’s most significant developments was the launch of Amazon Q, a generative AI-powered assistant that introduced natural language processing (NLP) capabilities to the BI market.

“One of the biggest challenges individuals face is knowing what to ask…Q now understands context, hints at follow-up questions, and provides multiple visual answers for advanced exploration,” Tracy pointed out.

Tracy described Amazon Q Business as a breakthrough tool for managing data at scale, with its ability to connect seamlessly to over 40 enterprise data reports, including Microsoft 365, Amazon S3, Google Drive, Salesforce CRM, and Asana.

The AI-powered assistant can synthesise data from various reports and provide individuals with actionable insights via natural language enquiries. “Amazon Q Business brings AI directly into the hands of business individuals to answer critical questions, automate key tasks, and generate visuals with ease,” Tracy mentioned, explaining how this allows teams to interact with data intuitively.

The productivity gains from Amazon Q have been substantial for businesses. , preliminary tests indicate that Amazon Q will increase staff productivity by as much as 80%, specifically through the automated extraction of insights.

However, he noted that the actual success measure is more than just the metrics; it is how successfully it is accepted and used across organisations.

The integration of generative AI to QuickSight was further amplified in its capability through the launch of scenarios analysis capability in Amazon Q in QuickSight.

“It’s a decision-making assistant,” Tracy explained. “people can simulate outcomes, and Amazon Q returns actionable insights and recommendations in real time.”.

“AI isn’t here to replace analysts. It’s here to make their work more strategic by automating repetitive tasks and uncovering insights they might otherwise miss,” Tracy mentioned.

As Amazon Q in QuickSight pushes the boundaries of what BI tools can do, it faces stiff competition from established players in the industry. Tracy believes Amazon Q in QuickSight’s native integration with AWS gives it a significant advantage.

“The integration with AWS services like data lakes, warehouses, and machine learning tools creates a secure, unified ecosystem for enterprises,” Tracy explained. This seamless connectivity not only enhances QuickSight’s functionality but also ensures that it remains a secure platform for business individuals, with stringent security measures in place to protect sensitive data.

A standout feature in Amazon QuickSight’s security capabilities is the Random Cut Forest (RCF) algorithm, which excels in real-time anomaly detection. “Unlike traditional machine learning algorithms, RCF is optimised for real-time anomaly detection, making it invaluable for fraud prevention and operational monitoring,” Tracy mentioned.

This focus on security underscores Amazon QuickSight’s commitment to safeguarding customer data while continuing to innovate with AI functions.

The popularity of Amazon Q in QuickSight can also be attached to its user-centric approach. “Most business intelligence technologies are built for specialists. We built Amazon Q in QuickSight to be self-service, where the intent is that every employee in the organisation can gain insights without requiring a broad degree in data science,” Tracy pointed out.

This shift from being a model that was data analyst-centric to making all those employees have their power has helped Amazon QuickSight acquire widespread adoption, considering hundreds of thousands of individuals depend on it daily.

Tracy imagines a world in which BI tools are woven into the fabric of all business processes. “Analytics should feel intuitive. It’s not just about visualising data – it’s about turning those visuals into actionable narratives that drive enhanced outcomes.”.

Tracy has valuable advice for aspiring BI professionals. “Focus on mastering AI-driven tools and developing a strong foundation in data storytelling. The ability to turn data into actionable narratives is what sets the best apart,” he mentioned.

Amazon Q in QuickSight changes how businesses engage with data by including elements such as scenarios and generative AI, making analytics a vital driver of strategic decision-making. “The future of BI is about making data accessible, actionable, and transformative for businesses of all sizes,” Tracy concluded.

Chennai-based data analytics business LatentView Analytics unveiled its financial results for the third quarter of FY25, reporting a total operating...

As an illustrative case study, we applied the framework to a dermatology model, which utilizes a convolutional neural network similar to that describe...

With DeepSeek becoming the world’s leading app in no time, ByteDance, the firm behind TikTok, has now released a research paper on its new video ge...