Of course, the state of the art here is still far from the scale and cost that would be necessary for practically useful devices, but I am optimistic about what can be realized in the next few years, especially given the potential these systems have to accelerate the very large-scale, very expensive AI systems like those used in popular textual ‘GPT’ systems like ChatGPT.”

“We don’t know what capabilities the next-generation ChatGPT will have if it is 100 times more powerful, but that’s the regime of discovery that this kind of technology can allow.”

“ChatGPT is limited in its size by the power of today’s supercomputers. It’s just not economically viable to train models that are much bigger. Our new technology could make it possible to leapfrog to machine-learning models that otherwise would not be reachable in the near future.”

Further, because the components of the system can be created using fabrication processes already in use today, “we expect that it could be scaled for commercial use in a few years. For example, the laser arrays involved are widely used in cell-phone face ID and data communication,”

opens an avenue to large-scale optoelectronic processors to accelerate machine-learning tasks from data centers to decentralized edge devices.” In other words, cellphones and other small devices could become capable of running programs that can currently only be computed at large data centers.

substantially several more orders of magnitude for future improvement.

With the new system, the team reports a greater than 100-fold improvement in energy efficiency and a 25-fold improvement in compute density, a measure of the power of a system, over state-of-the-art digital computers for machine learning.

In the July 17 issue of Nature Photonics, the researchers report the first experimental demonstration of the new system, which performs its computations based on the movement of light, rather than electrons, using hundreds of micron-scale lasers.

MIT-led

global shortage of GPUs

With the new KL730, Kneron has made some dramatic performance improvements to its NPU silicon. Liu said that the KL703 has better performance than prior generations and can also be clustered. As such, if a single chip isn’t enough for a specific use case, multiple KL703s can be clustered together in a larger deployment.

“The reconfigurable AI architecture can dynamically change the structure inside the chip to support almost any kind of new model,” Liu said.

The KL730 isn’t Kneron’s first chip optimized for transformers — the company announced the KL530 silicon two years ago, which had that capability.

“GPU hardware was specifically designed for gaming, and right now it’s just Nvidia trying to brainwash all of us trying to say that only a GPU can do AI,” said Liu.

neural processing unit (NPU)

The KL730 achieves an impressive 3-4x leap in energy efficiency compared to its predecessors and claims to be up to 2x more energy efficient than major competitors in the industry.

Yet Global Illumination’s name and previous work appear to be focused on visual-facing products and services, indicating that OpenAI may be pursuing more multimedia features for ChatGPT, or building out its Dall-E 2 image generation service, or perhaps launching a video-generation product to rival fellow New York City startup Runway with its Gen-2 text-to-video generator.

OpenAI says that the talent from Global Illumination will “work on our core products including ChatGPT.”

Global Illumination, Inc. team “have also made significant contributions at YouTube, Google, Pixar, Riot Games and other notable companies.”

Meta

Gartner Hype Cycle

Gartner Hype Cycle places generative AI on the ‘Peak of Inflated Expectations’

LimitationsJudgments by language models are vulnerable to undesired biases that might have been introduced into the model during training. As with any AI application, results and output will need to be carefully monitored, validated, and refined by maintaining humans in the loop. By reducing human involvement in some parts of the moderation process that can be handled by language models, human resources can be more focused on addressing the complex edge cases most needed for policy refinement. As we continue to refine and develop this method, we remain committed to transparency and will continue to share our learnings and progress with the community.

With this system, the process of developing and customizing content policies is trimmed down from months to hours.

A content moderation system using GPT-4 results in much faster iteration on policy changes, reducing the cycle from months to hours.

A noteworthy irony There is of course a noteworthy irony to OpenAI’s promotion of GPT-4 as a way to ease the mental burden of human content moderators: according to detailed investigative reports published in Time magazine and The Wall Street Journal, OpenAI itself employed human content moderators in Kenya through contractors and subcontractors such as Sama, to read content, including AI-generated content, and label it according to the severity of the severity of the harms described. As Time reported, these human laborers were paid less than $2 (USD) per hour for their work, and both reports indicate that workers experienced lasting trauma and mental illness from it. “One Sama worker tasked with reading and labeling text for OpenAI told Time he suffered from recurring visions after reading a graphic description of a man having sex with a dog in the presence of a young child,” the Time article states. Workers recently petitioned the government of Kenya to enact new laws that would further protect and provide for content moderators. Perhaps then, OpenAI’s automated content moderation push is in some sense, a way of making amends or preventing future harms like the ones that were involved in its creation.

Interestingly, OpenAI’s blog post also took the time to call out the “Constitutional AI” framework espoused by rival Anthropic for its Claude and Claude 2 LLMs, in which an AI is trained to follow a single human-derived ethical framework in all of its responses.

Media organizations have long struggled with effectively moderating reader comments on articles, while still allowing for freedom of speech, discussion and debate.

How GPT-4’s content moderation works

In fact, according to OpenAI’s research, GPT-4 trained for content moderation performs better than human moderators with minimal training, although both are still outperformed by highly trained and experienced human mods. <img data-lazy-fallback="1" decoding="async" width="1292" height="844" src="https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?resize=1292%2C844&#038;strip=all" alt="" class="wp-image-2893565" srcset="https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=1292&amp;strip=all 1292w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=300&amp;strip=all 300w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=768&amp;strip=all 768w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=800&amp;strip=all 800w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=400&amp;strip=all 400w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=750&amp;strip=all 750w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=578&amp;strip=all 578w, https://venturebeat.com/wp-content/uploads/2023/08/Screenshot-2023-08-15-at-4.40.33-PM.png?w=930&amp;strip=all 930w" sizes="(max-width: 1000px) 100vw, 1000px" data-recalc-dims="1" />Credit: OpenAI

The content they are exposed to often includes detailed descriptions and photographic or video evidence of humanity at its worst — such as depictions of child sexual abuse — not to mention various other crimes, atrocities and horrors. Moderators charged with identifying and removing this content have reported struggling with post-traumatic stress disorder (PTSD), anxiety and various other mental illnesses and psychological maladies due to their exposure.