Intel just dropped the most powerful 7B model ever!

- Top 7B on Huggingface's leaderboard.

- Base model: Mistral 7B.
- Fine Tuned on: SlimOrca
- DPO: LLaMA-13B vs ChatGPT Gens (Prefer ChatGPT).

Unique Hardware:

Trained on Habana Labs's 8x Gaudi2 mezzanine accelerators.

- Comparible in performance to A100.

- 96GB on device RAM.

Hong Kong investment firm Victory Securities obtains retail crypto trading license

Victory Securities said it is the first licensed corporation to offer a crypto trading and advisory service to retail investors. It joins crypto-native firms HashKey Exchange and OSL Digital Securities.

Hong Kong introduced a regulatory regime permitting the offering of crypto services to retail investors earlier this year, in a sign of the cryptocurrency market shifting eastward as Asian jurisdictions offer greater clarity to companies than they may find elsewhere, such as in the U.S.

Customer support productivity rises 14% with LLM AI tool, with a 35% boost for the least experienced workers.

⚡️ OpenAI has hired the former head of Google's TPU AI chip efforts as head of hardware, and is hiring for "experts in data center facility design" to help inform its partners (and maybe make its own chips).

New report from Deloitte various business challenges and importantly the components that are ready now - scalability, security, developer tools, financial applications, digital assets and enterprise blockchain.

This AI system identifies animal sounds, tracks wildlife with camera traps, and models ecosystem changes.

Its data-driven approach is a game-changer for understanding and protecting biodiversity.

TSMC Chairman Mark Liu said the world’s semiconductor industry faces a slowdown in innovation due to the US government’s ‘AI Sanctions’ (export curbs) on chip technology to China

the comments were made in a speech about AI.

As China chip makers are forced to focus on R&D that’s old hat elsewhere, while reduced competition and the loss of China revenue hurts chip R&D outside of China.

He fears globalization has been hurt, and could be split by geopolitical tension.

TSMC’s chairman also predicted Nvidia will become the biggest semiconductor firm in the world in 2023 amid the rapid growth in AI, noting growth among fabless companies (chip designers) is increasing and will be 10% (CAGR) average growth over the next 5-years, versus 4% for IDMs (Integrated Device Manufacturers – which design and manufacture, like Intel).

Our brains have an amazing ability to 'rewire' themselves after loss of sight, amputation or stroke... right?

Wrong, say Professor Tamar Makin and John Krakauer.

They say that what is occurring is merely the brain being trained to utilise already existing – but latent – abilities.

The newest industrial PC VR/AR headset by Varjo - the XR-4 - promises passthrough camera quality so good that it blends with the retina-level visuals…and has an optional auto-focus camera.

⚡️ Rumor is that Q-Star figured out a way to break encryption, and OpenAI tried to warn the NSA about it.

Here’s a Google doc link to a compilation of (allegedly) leaked documents and compelling analysis.

Vitalik Buterin stated at the Devconnect in Turkey that he plans to redesign Ethereum staking and solve problems affecting performance.

Buterin recognizes the UTXO payment model and hopes to integrate private mempools, ERC-4337, code pre-compilation, ZK-EVMs and liquid staking in Ethereum.

He also expressed concern about the increasing concentration of Ethereum liquid staking in the proof-of-stake model.

New neurotech eschews Electricity for Ultrasound: Forest Neurotech will use ultrasound-on-chip tech from Butterfly Network to develop a brain computer interface.

The Spanish Tax Administration has issued a tax form, requiring Spanish citizens to declare their crypto-assets held on overseas cryptocurrency platforms when an individual’s balance sheet exceeds $55,000 in crypto-assets.

Training of 1-Trillion Parameter Scientific AI Begins

A US national lab has started training a massive AI brain that could ultimately become the must-have computing resource for scientific researchers.

Argonne National Laboratory (ANL) is creating a generative AI model called AuroraGPT and is pouring a giant mass of scientific information into creating the brain.

The model is being trained on its Aurora supercomputer, which delivers more than an half an exaflop performance at ANL.

The system has Intel’s Ponte Vecchio GPUs, which provide the main computing power.

Intel and ANL are partnering with other labs in the US and worldwide to make scientific AI a reality.

“It combines all the text, codes, specific scientific results, papers, into the model that science can use to speed up research,” said Ogi Brkic, vice president and general manager for data center and HPC solutions, in a press briefing.

Brkic called the model “ScienceGPT,” indicating it will have a chatbot interface, and researchers can submit questions and get responses.

Chatbots could help in a wide range of scientific research, including biology, cancer research, and climate change.

Pika Labs now generates video from text, expands video canvas size, and edits moving objects from just a prompt.

Crazy to think anyone will be able to generate entire movies from their phones soon.

GenAI could transform how health care works

I. Technology Substitution vs. Ecosystem Transformation

By combining and analyzing data acrosss previously disconnected silos, generative AI creates the opportunity to raise the bar on efficiency and effectiveness across the spectrum of health care delivery

1. Billing and Claims
Allowing artificial intelligence to break the silos between insurers, hospitals, and consumers would automate claims management, prior authorization, and even payment planning and collections, helping to eliminate a massive drag on system efficiency

2. Resource Management
AI will enable cross-platform coordination across hospitals, systems, partners, and vendors to create higher resilience and better patient placement, lowering risk, shorten recovery times while improving outcomes and lowering cost

3. Redefining Quality
By incorporating the latest advances in medical science and real-world evidence into treatment recommendations and measures, AI stands to improve patient outcomes and raise standards in ways that reduce burden on both the patients and the system.

II. Ecosystem Transformation Requires Organizational Transformation

1. Changing data access changes authority
This is a huge ecosystem transformation, shifting the focus from insuring the accuracy of content (“Is the data correct?”) to controlling the breadth of questions (“Who is allowed to ask what?”).

2. New information demands new metrics
New visibility into new data combinations open debates on relevant and appropriate metrics which, in turn, impact goals and incentives

3. Transparency creates new responsibility
A corollary to visibility across data silos is the expectation of more holistic decisions that take the broader landscape into consideration

III. Changing Asymmetries and Strategy-Making

1. When scale can be aggregated, size matters less
The world of AI creates potential for decentralizing — and more equitably distributing — the delivery of care by increasing the viability of smaller institutions could be a surprising upshot of the current revolution

2. New connections drive new synergies
A truly holistic view of the patient as a person — health, employment, living situation, social needs — offers an opportunity to redefine care and delivery by revisiting the organization of activities across the system, lat long last manifesting the promise of profitable value-based care

3. Winners take action — and action transforms the game

Microsoft Chief Scientific Officer Eric Horvitz explains how new prompting strategies can enable generalist large language models like GPT-4 to achieve exceptional expertise in specific domains, like medicine, and outperform fine-tuned specialist models.

Amazon has announced Q, a new AI chatbot for businesses

The chatbot helps corporate users with tasks like summarizing reports or answering policy questions.

Q aims to attract users weary of consumer chatbots handling sensitive info.

Amazon AWS also announced a host of new partnerships with NVIDIA, including:

-NVIDIA BioNeMo — a generative AI platform for drug discovery is coming to AWS.
-AWS becomes the first cloud provider to deploy GH200 chips.

Lately Epic announced that it would tap Microsoft to accelerate generative AI-powered tools to help clinicians save time while using their electronic medical record systems.

Now Sumit Rana, head of research and development, gave an interview about that. A few take-aways:

1. Physicians ask the patient at the beginning of the visit if they can record the visit to help with documentation, and most patients are okay with it.

2. And then once that conversation is done, a few moments later, the note gets updated with a draft of what was discussed, and then a physician can review that note, make changes similar to what they might do in a dictation workflow, and then they finalize their documentation.

3. One site reported average savings of five-and-a-half hours per week. Another one looked at time spent by doctors after clinical hours, and they saw a 76% reduction in that time spent after clinic hours.

4. The other big learning we had was that it's super important for the provider to be able to hover and see citations so they can see not only what the summary is, but what facts within the medical record were used to generate this summary.