Scientists have discovered a previously unknown mechanism by which cells break down proteins that are no longer needed

Their discovery potentially provides a new pathway that could be useful for tackling many diseases including cancer.

It’s a example of the important research that AlphaFold is helping to enable.

Huawei graphics chips already match Nvidia’s A100 in terms of capability, and Huawei will soon take the dominant position in computing power away from Nvidia, media report, citing the chairman of China’s iFlytek, which is poised to launch a GPT-4 rival.

DeepMind launched 𝗦𝘆𝗻𝘁𝗵𝗜𝗗: a digital tool to watermark and identify AI-generated images.

SynthID can embed a digital watermark directly into the pixels of AI-generated images, making it imperceptible to the human eye.

By scanning an image for the watermark, SynthID can assess the likelihood of it being created on Imagen.

Tesla is about to flip the switch on its new AI cluster, featuring 10,000 Nvidia H100 compute GPUs.

Tesla launched its highly-anticipated supercomputer yesterday. The machine will be used for various AI applications, but the cluster is so powerful that it could also be used for demanding high-performance computing (HPC) workloads.

In fact, the Nvidia H100-based supercomputer will be one of the most powerful machines in the world.

Tesla's new cluster will employ 10,000 Nvidia H100 compute GPUs, which will offer a peak performance of 340 FP64 PFLOPS for technical computing and 39.58 INT8 ExaFLOPS for AI applications. In fact, Tesla’s 340 FP64 PFLOPS is higher than 304 FP64 PFLOPS offered by Leonardo, the world’s fourth highest-performing supercomputer.

With its new supercomputer, Tesla is significantly enhancing its computing capabilities to train its full self-driving (FSD) technology faster than ever. This could not only make Tesla more competitive than other automakers but will make the company the owner of one of the world's fastest supercomputers.

The scaling hypothesis of LLMs in biology - to what degree increasing the size of a model and training set improves performance and enables general purpose algorithms - will be the most interesting experiment in computational biology of the next decade.

This will require a better understanding of what protein language models learn about the underlying biology (eg sequence-structure-function relationships), and overcoming the context limitation of training LLMs on a single biological data type in order to capture the broader cellular/physiological context of a biological system.

From Concussion to Alzheimer’s: Mini-Brain Unlocks Time Link
Researchers developed a “TBI on a Chip” to explore links between TBI and Alzheimer’s

Study:
- Researchers exposed the functioning clusters of cultured neurons from mice
- to 3 blows of 200 g-force,
- each approximating the higher end of what a football player receives in a single hit

- Using a device with a pendulum to deliver a specific g-force 
- to a chamber with a cluster of a quarter million neurons.

Key Findings:
1) Within the first 24 hours after a hit:
- 350% increase in the production of misfolded AB42,
- linking immediate effects of TBI to long-term alzheimers risk.

2) The device utilizes a microelectronic array to measure neuronal electrical activity

3) Hydralazine (an FDA-approved drug for lowering blood pressure)
was found effective in reducing the levels of acrolein and misfolded AB42.

“This amyloid beta pathology started within hours, maybe immediately. That’s never been heard of,” Shi said.
“It’s like attacking the weight-bearing stud in a house wall. If you break that stud, ... the house is going to fall down.”

Future directions:
The device can also be used to test possible therapeutics, including drugs known to reduce acrolein levels.

The biggest surprise at HotChips is Numenta.

Although some of the tech details were undisclosed, they seem to exploit sparsity in SW in extreme to beat CPUs in low batch inference.

They are 3x faster on 48 client one shot BERT inference vs 8-batch A100.

This was achieved with Xeon MAX w/AMX & HBM but obviously more memory bound than compute.

Faster than Nvidia H100

Habana Labs Gaudi2 achieves up to a x1.4 speedup compared to H100 for training a vision-language model like BridgeTower.

Baidu among first cos. to get Beijing's nod to roll out their AI models to the public.

The approvals officially bring ChatGPT-style services to one billion or more internet users for the first time.

Ready Player Me launches AI-based Copilot to help developers streamline avatars

Ready Player Me a provider of cross-platform avatar platforms, is introducing Copilot, a new tool fueled by OpenAI’s GPT-4 and OpenCopilot.

Designed to assist developers in seamlessly integrating the avatar platform into their applications and games, Copilot leverages machine learning and extensive training on technical documentation and support tickets.

Ready Player Me offers a range of free tools, including Unity and Unreal Engine SDKs, as well as web integration tools.

By integrating Ready Player Me’s platform, companies can save valuable development resources that would have been spent on creating their own avatar system from scratch. This integration streamlines the process, saving both time and money.

Open source AI grants from a16z

Fasikl announced that it has received two simultaneous breakthrough device designations from the U.S. FDA for its minimally invasive implantable platform:

The MindForceTM NCI System is a nerve-computer interface system that enables amputees and paralyzed patients to effortlessly control a robotic hand or computer device, with precise control over individual finger joints and allows them to be able to do everyday tasks, simply by thinking.

The EpioneTM system effectively alleviates chronic, intractable postamputation pain by recreating sensations through peripheral neurostimulation.

All 4 projects that Sam Altman is most involved in have raised 9+ figures before releasing their first commercial product:

• OpenAI (~1B*+)
• Helion Energy ($580M)
• Retro Bio ($180M)
• Worldcoin ($125M)

How can we increase both the cultural and VC appetite for megaprojects?

Jais is a bi-lingual GPT model trained on our recently-announced Condor Galaxy 1 AI supercomputer, with a unique combination of Arabic, English, and code.

It is a product of strategic partnership with G42, co-developed by G42's Inception, MBZUAI (Mohamed bin Zayed University of Artificial Intelligence), and Cerebras Systems.

The model is open-source posted on Hugging Face with a permissive Apache 2.0 license for research and commercial use, to better engage and enable Arabic and multi-lingual communities of users, researchers, open and enterprise application developers.

This is a neural network flying a drone at extremely high speed, beating human champions in FPV drone racing.

- Reinforcement learning as a tool is so marvelously versatile. It's able to solve both fast, reactive tasks and slow, deliberate tasks (ChatGPT RLHF).
- Trained in large-scale simulation, finetuned in real world - I believe this is the paradigm that will get us to generalist robot some day.

This is amazing. PanoHead is a new model that generates 3D textured models from a single image.

Github: sizhean.github.io/panohead

Meta AI announced Belebele, a first-of-its-kind multilingual reading comprehension dataset. This dataset is parallel for 122 language variants, enabling direct comparison of how well models understand different languages.

Samsung reveals the world's first 32 Gb DDR5 DRAM die, allowing the company to lower costs of high-capacity memory modules and build 1 TB RDIMMs for servers.

What we think is intelligence might not be the intelligence we see in nature.

“This experiment of a fish swimming upstream is truly stunning. Why? The fish is dead. If you’re streamlined and flexible, you can do a lot by doing nothing at all,” - Christian Hubicki.

Before we reach Human-Level AI, we have to reach Fish-Level AI.

A biotech company says it put dopamine-making cells into people’s brains

In an important test for stem-cell medicine, biotech company BlueRock Therapeutics says implants of lab-made neurons introduced into the brains of 12 people with Parkinson’s disease appear to be safe and may have reduced symptoms for some of them.

How it works: The new cells produce the neurotransmitter dopamine, a shortage of which is what produces the devastating symptoms of Parkinson’s, including problems moving. The replacement neurons were manufactured using powerful stem cells originally sourced from a human embryo created using an in vitro fertilization procedure.

Why it matters: The small-scale trial is one of the largest and most costly tests yet of embryonic-stem-cell technology, the controversial and much-hyped approach of using stem cells taken from IVF embryos to produce replacement tissue and body parts.

From rocket engines to software for the entire engineering industry

Flow, an ex-rocket company turned software, wants to change that on episode 7 of S³.