Turning scaling into a systematic science is the biggest advance enabled by LLMs.

AI search of Neanderthal proteins resurrects ‘extinct’ antibiotics

Bioengineers have used AI to bring molecules back from the dead. These newly identified protein snippets from Neanderthals are found to have bacteria-fighting powers.

Andrej Karpathy, co-founder of OpenAI: Today FLOPS is one of the things you can spend $ on.
Tomorrow $ is one of the things you can spend FLOPS on.

Flow2 - a revolutionary breakthrough in neuroimaging

Flow2 offers fMRI-like functional neuroimaging through TD-fNIRS + EEG integration. Portable & built to scale - let's unlock the potential of precision neuromedicine together!

How exactly the FATF travel rule for crypto-assets will apply to the Bitcoin  lightning network is still unclear.

However, the JMLSG in the UK has now published guidance that could become highly influential, not only in the UK, but around the world.

It says lightning transfers are in scope of the travel rule, but only the ultimate originator and beneficiary of the transaction, not the intermediary (routing) nodes.

You can give feedback on this approach until August 25th: https://www.jmlsg.org.uk/consultations/consultation-cryptoasset-transfers/

This guidance is issued by a private sector body and therefore not legally binding, but it has HM Treasury approval and therefore is very significant. It is rather high-level at this point, but more detailed policies and procedures will be based on it

The Monetary Authority of Singapore (MAS) today announced that it will commit up to S$150 million over three years including innovative FinTech solutions arising from emerging technologies such as Web 3.0, capped at S$2 million per project.

Backend for enterprises. Multiplayer raises $3M for AI-based distributed software development.

Apple stands to save $ billions from a deal for TSMC to pay for all defective 3nm Apple A17 chips

Note: Doubt Apple would save "$ billions", as it sounds like each side can tell their boss they negotiated great terms. Apple-TSMC have a symbiotic relationship, so if $$ pain is too much for either, it ultimately hurts both.

Researchers found that human cortical spiking organizes into a backbone sequence that persists over time, and cognition involves rearrangement around this backbone.

Sense Arena unveils VR-based NHL ice hockey training with sensors on sticks

Sense Arena is launching a virtual reality game for ice hockey training dubbed NHL Sense Arena for the Meta Quest 2 headset.

The game is coming this fall, and Sense Arena has a multi-year agreement with the NHL for the noscript. Sense Arena has been out in the market since 2018, but this is the first experience that carries the NHL brand name.

With the game, you can strap on sensors including an Oculus Quest hand controller to a hockey stick.

Then, while wearing VR goggles that immerse you in a hockey rink, you can swing at virtual pucks with the real stick and the game will translate your movements into the game. The video shows what that experience will be like, and it looks like it has pretty good precision and it requires skill.

This platform will allow hockey enthusiasts to refine their skills, compete, and experience what it’s like to be a part of their favorite NHL franchise within the Sense Arena hockey environment. NHL Sense Arena will offer players and fans an immersive experience by featuring the uniforms and logos of all 32 NHL teams.

NVIDIA DGX Cloud will be accessible with HuggingFace to create & customize generative AI models for the enterprise.

- Each node is either 8x H100 or A100.
- NVIDIA nForce3 Pro will be the networking technology that powers massive scaling.
- NVIDIA Base Command will be the OS for DGX Cloud. It includes a cluster API.

Google announces a new browser-based code environment.

It will bring the entire full-stack and app development workflow to the cloud.

It also includes Generative AI features based on PALM 2:
1. Code generation
2. Code completion
3. Translating code between languages
4. Code Explanation

You can join the waitlist today.

IBM and NASA deploy open-source geospatial AI foundation model on Hugging Face

IBM and NASA have announced the release of the watsonx.ai geospatial foundation model on Hugging Face, expanding the collection of open-source models available.

The model, developed to harness the potential of satellite imagery for climate science and real-world improvements, was trained using NASA's Harmonized Landsat Sentinel-2 satellite data.

It has been fine-tuned for specific applications such as burn scar and flood mapping.

Anthropic introduced a latest version of Claude Instant, version 1.2

Claude Instant 1.2 is a better at math/code, languages, structured outputs & formatting. And very fast.

Anthropic also made improvements on safety — less hallucinations and more resistant to jailbreaks. That said, solving honesty is a new dream for thinking machines.

Great article covering OpenBCI's partnership with Lenovo. Human-computer interaction has progressed from punch cards, to keyboards, to trackpads and touchscreens.

What’s next...?

"Within a decade, PC users could start having a lot more intuitive computing experiences. Based on data from technology like Galea and other sources, as well as new AI-enabled applications, computers will understand, predict and respond to individual’s needs based on biometric data and other inputs."

“Ultimately, I see the combination of neurotechnology and mixed reality as the future of personal computers,” says Russomanno. “Spatial computing, BCI, and AI will converge this decade, resulting in quite possibly the greatest technological inflection point humanity ever experienced. It is going to be remarkable.”

How will we look back on the current era of HCI innovation?

IBM announced it will host MetaAI's Llama 2-chat 70 billion parameter model in the watsonx․ai studio

The Worldcoin authority that submits zk proofs of eyeball scans spent $500k in gas over 40 days.

$4.5 million annual in gas fees.

Companies will need to hoard ETH the same way they hoard other essential commodities.

Making AMD GPUs competitive for LLM inference. Worried about NVIDIA GPU shortage?

This project aims to makes it possible to compile LLMs and deploy them on AMD GPUs using ROCm and get competitive performance.

Discussion here.