6ai Focused Language Models | FLMs
More focused and More Sustainable
Let’s look at the models that will be best suited for the future of work and humanity. Those that can be more sustainable and practical, and everyone can use. But first, let’s examine the current state of Large Language Models (LLM) and Generative AI. Scientist Gary Marcus, a leading voice on AI and best-selling author and entrepreneur. Is on record way over ten years ago anticipating many of the current limitations and challenges AI is now experiencing.
Prof. Marcus’s research in human language development and cognitive neuroscience challenges the delusions and non-truths about AI. Dave Ferrucci, an award-winning AI scientist who started and led the IBM Watson team, is also aligned with Marcus about the limitations and realities surrounding AI.
Both agree that ‘deep learning’ is an incredible invention that is great for pattern recognition, making ‘predictions’ and extrapolating outputs from data. However, what it’s fundamentally lacking is the ability to understand why it produces the answers it does. Unlike humans, AI is not conscious about what it’s doing and can’t ever build any understanding or authentic reasoning into its answers. LLM has no embedded mechanisms for background understanding of its text/outputs. While humans seamlessly have this critical function.
Neural networks have trouble generalizing beyond the multidimensional space that surrounds their training examples. That limits their ability to reason and plan reliably. It also drives their greediness with data and even the ethical choices their developers have been making. There will never be enough data; there will always be outliers. This is why driverless cars are still just demos, and why LLMs will never be reliable.
— Gary Marcus
So where do we go from here? To first principles, of course. What Einstein has taught us about thinking in reality — the six-step applied intelligence (6ai) is based on Einstein’s process of discovery and logic. These principles adhere to reality and fill the gap with a useful process methodology for business problem-solving. These six principles seek Insight through objective truth!
6ai will not get trapped in the abyss of Platonism — abstract concepts and theories — becoming a distraction more than anything else. Instead, ai seeks to serve and deal with only what is real, practical, and can be effectively applied. Theories are useless if they can’t be effectively applied.
So solutions focused on the utilization and deployment of effective tools for humans to build intelligent strategies to enhance human innovation is what ai is about. By combining the value functions of artificial intelligence (AI) with those of human applied intelligence (ai) we can develop highly focused and effective language models to gather insights for strategy development.
Our Focused Language Models (FLM) connect with existing relevant knowledge bases, with template applications, with speed and accuracy. Retrieving quality information at a higher dimensional level of relevant content. Enabling FLMs to have access to up-to-date, well-sourced, relevant information that can be transmitted and applied effectively.
The six-step (6ai) process guides users through a defined scientific strategy development process, retrieving highly relevant information from verified and reputable sources that can be taken as empirically warrantable.
The retriever and reader method enables the FLMs to query corpuses of data and the open-source frameworks available, making it easy to build customized FLM templates for useful insight generation.
Focused Language Models (FLM) are purposeful retrieval-augmented-generation methods, that give models access to highly relevant knowledge bases. Which are also more practical, efficient, and considerably less expensive than training LLMs from scratch. And much easier to implement and manage.
Next, the 6ai strategy canvas allows for the designing of strategy. Anything from high-level multi-billion-dollar corporate strategies to personal growth strategies. Anyone can do it. And at an infinitesimal fraction of the cost of hiring useless management consultants.
6ai Technologies FLMs have been specifically crafted for simplicity and accessibility for all! It’s rapidly democratizing access to technical capabilities once dominated by consultants and advisors acting as courtiers for the Big Tech aristocracy.
It’s about who controls the technology that matters.
6ai utilizes GenAI purposefully and responsibly to augment human capacity and ingenuity, which leads to creating extraordinary value for organizations and individuals. It inherently works to serve humanity best.
In a survey (Harvard Business Review) 10, 000 senior leaders:
- 97% said strategy is the most important thing to their organization’s success, and 96% said they lack the time and the right tools to effectively engage in strategy development.
When executives were asked by the MIT Technology Review Insights Survey about the most important business objectives they have set over the next two years, the vast majority responded: substantial degree of alignment with a growth-oriented business strategy.
The main problem companies have identified is the lack of scalable, no-nonsense, and easy-to-use digital solutions with a central insight resource. Existing GenAI solutions, i.e., ChatGPT are too error-prone and unreliable to fill that gap.
Respondents also said that finding new platform organizational solutions that enable enterprise-wide team collaboration, and data-driven strategy development, is priority #1.
Architectures have gotten really complicated, but only because we tend to over-complicate them. We do this because we lose sight of what matters most. We too often bring in the latest and greatest in technology, thinking they will solve the problem. But unless the business is ready to leverage the tools, and has the maturity to extract the insights and processes and agreed-upon logic. We’re only adding to the spaghetti architecture.
— Sol Rashidi, chief analytics officer at The Estée Lauder Companies
Language Processing (NLP) or Large Language Models (LLMs) does offer some cool things, but not what is practically needed. They become distractions.
The computational demands necessary for LLMs to be focused are not there, so it makes it impractical to build highly focused and specific strategy solutions templates. If for example, a clinician wants to develop a strategy for a Black mental health therapy program that specifically addresses the societal trauma experienced by Black people. A significant need! You would be hard-pressed to find anything useful in the existing data, if any at all, which in itself is part of the problem.
6ai’s FLM-template models are built to be hyper-focused on the strategy and insight demands of the user; with 21st Century application and delivery capabilities.
Tremendous demand exists for specialized or smaller-scale models, or tailored solutions that are fit for purpose. Society can’t progress and get anything done if the existing solutions engage them in an infinite loop of hallucinations.
FLMs align precisely with the expressed strategy demands of the users. This must be the number one thing, users need, not what Big Tech wants to sell them.
In a report titled The Great Acceleration: CIO Perspectives on Generative AI; by MIT Technology Review. The report was based on in-depth interviews with senior executives and experts, conducted in April and May 2023. It seeks to understand how technology leaders are adopting emerging generative AI tools as part of building enterprise-wide, AI-driven strategies.
The report also draws on MIT’s global survey of 600 senior data and technology executives, conducted in May and June 2022.
A major part of the report focuses on the customization of language models to meet specific industry needs, please see below.
Before generative AI, few organizations had adopted AI as a critical part of any business function. Now, the opportunity and the technology exist for organizations to demand language models that fit their purpose. This is 6ai’s mission. Organizations are the customers so they must push for what they want, and lock the gates to the zombie consultants that are coming to sell them the Big Tech agenda.
If you care deeply about a particular problem or you’re going to build a system that is very core for your business, it’s a question of who owns your IP, says the report. DuPont is a science and innovation company, and there is a need to keep LLM models internal to our organization, to protect and secure our intellectual property — this is a critical need.
— MIT Technology Review Insights.
From an overall security and competitive advantage perspective, you don’t want an existing LLM where the data that you’re putting in could be used by others to compete against you.
The report also highlights that some CIOs are not waiting and taking steps to limit company use of external generative AI platforms. Samsung, for example, has banned ChatGPT after employees used it to work on commercially sensitive code. A raft of other companies, including JP Morgan Chase, Amazon, and Verizon have enacted restrictions or bans.
We can’t allow things like Shell’s corporate strategy to be flowing through ChatGPT, says a Shell executive.
Since LLMs harness the totality of the online universe, well-informed companies will become more cautious about what they put online and seek to develop their own strategies internally. Free of consultants who only ‘copy and paste’ and will certainly put it into their databases, to generate new business and share with their clients.
MIT says that inaccurate and unreliable outputs are a bigger worry because they can develop flawed strategies that can lead to meaningful business underperformance. Creating unnecessary dangers for the enterprise. Who needs that?
The largest LLMs are, by dint of their size, tainted by false information online, says Matei Zaharia, co-founder and chief technology officer at Databricks and associate professor of computer science at the University of California, Berkeley.
He goes on to say that winning organizations should be doing things…in a more focused domain, and you can avoid all the random junk and unwanted information from the web.
The applied intelligence six principles/steps provide the most practical and effective development system for FLMs. It precisely aligns with the organizations’ objectives, not with Big Tech’s. MIT says that
- Smaller models provide a viable alternative,
- The reduction in complexity comes by narrowing your focus from an all-purpose model that knows all of the human knowledge to very high-quality knowledge just for you because this is what individuals in businesses need
Thankfully, smaller does not mean weaker. Less is more for 6ai/FLMs. Generative AI models can be fine-tuned for domains requiring less or more concentrated data, as evidenced through models like BERT — for biomedical content (BioBERT), legal content (Legal- BERT), and French text (the delightfully named CamemBERT).
- Smaller open-source models could rival the performance of large models and allow practitioners to innovate, share, and collaborate. One team built an LLM using the weights from LLaMA at a cost of less than $600, compared to the $100 million involved in training GPT-4.
- All the large models that you can get from third-party providers are trained on data from the web. But within your organization, you have a lot of internal concepts and data that these models won’t know about.
Therefore, 6ai FLMs combines both internalized and externalized data, including anonymized data into a simple one-stop strategy template solution, specific to the user’s domain and strategy development needs.
6ai FLMs puts world-class strategy development in the hands of people and organizations who want to be in control of their future. And do not want to pay the high costs of consultants who just end up telling you what you already know.
It’s not just the OpenAIs/Microsoft and the Googles of the world that can lead the future. We the people can too! And 6ai exists to empower ordinary people to build extraordinary strategies.
Generative AI is increasingly democratizing access to technical capabilities which was previously confined to Big Tech and the Big Consulting industry. However, people can craft their own strategies and manage their organizations once given the right tools.
MIT Insights predicts “that ideas for AI will start coming from the workforce, marking the start of a more self-service and entrepreneurial era within organizations.” 6ai exists to facilitate that future.
6ai sees a world where FLMs largely replace LLMs; allowing anyone from executives to students to find high-quality data-driven insights, and build and own their strategies. Contributing to the environment by putting aside high energy consumption LLMs/businesses like ChatGPT.
There is a world where we can earn well and do good simultaneously. These are not mutually exclusive concepts by any means.
Right now, generative AI is more hype than anything else. 6ai doesn’t buy into the hype or the hallucinations. We see useful, responsible, purposeful and practical applications, via FLMs. Helping humanity accelerate innovation and productivity with optimal principles for building good strategies for winning.
