Jem Davies

Take the twenty minutes and listen to James Cameron´s view on the development of AI and robotics. Quite interesting! https://lnkd.in/eg9smzQv #jamescameron #ai #robotics

1

How much trouble can a 5-digit code cause? Well, if the code in question is part of the UK Standard Industrial Classification (UK SIC) scheme, then it turns out the answer is....quite a lot! UK SIC codes are maintained by the Office for National Statistics (ONS), as "The system used in classifying business establishments and other statistical units by the type of economic activity in which they are engaged.". The reason why UK SIC codes are important is that they provide the underlying 'model' that the Government considers when thinking about supporting industry sectors. UK SIC codes are the mechanism by which Government recognises that an industry exists and the kinds of economic activity it supports. Currently, the complexity of the UK video games sector is defined by two numbers: - 58.21/0 Publishing of video games (Section J), and; - 62.01/1 - Ready-made interactive leisure and entertainment software (Section K Information and Communication) These codes were set up in a world in which video games arguably did break down quite neatly into 'publishers' and 'developers'. But that is not the world we live in today. The problem now is that because of the inaccuracy of these two codes as a means of representing our industry, Government and other industry data which draws from them is getting messier and messier. Some developers, frustrated with these definitions, have opted to use other codes elsewhere in the system relating to software development. Some look to codes that refer to the creation of artistic works. Data provided by Ukie in 2016 suggested that as many as 59% of UK of games businesses were classified under the wrong code - and therefore effectively invisible to policymakers. Games are one of the UK's very few true growth industries. We occupy a unique vertical at the intersection of creativity, technology and community and can't easily be classified into any single one of these. By failing to represent the nature of our industry, SIC codes risk acting as a constraint on our economic ambitions for growth - because we are inaccurately and incompletely reflected in official data and statistics (and therefore in calculations about how games businesses should be supported). So why not just correct the system? SIC codes are a global classification Scheme and like all such endeavours naturally resistant to change (foundational standards like this often are, since the implications of changes echo through entire industries). We could replace UK SIC codes with something that works for our industry - but this would require DCMS and others to abandon a mandatory framework for an industry-defined one. So our best option is to lobby for political support to change the SIC codes for video games and interactive entertainment as a matter of urgency. In a very real sense the Government's ambitions for growth hinge on having the right underlying structures to identify and support video games as a leading Creative Industry.

22

10 Comments

Shares in Raspberry Pi jumped 39% when they began ‘conditional trading’ on the London Stock Exchange’s Main Market this morning as part of the low-cost computer maker’s #IPO. It’s a promising start to listed life for Raspberry Pi, with unconditional trading for the public set to begin on Friday 14 June and high levels of interest in this rare UK tech IPO. Raspberry Pi priced its shares at 280p on IPO, giving a market capitalisation of £514.6m on commencement of conditional dealings. When trading began at 8am this morning shares popped to 392p, before settling at around 367p as I write. The initial offer size of £166m equates to around 31% of the company’s ordinary shares on admission, however if, as seems likely, an over-allotment option is exercised, then the final offer size will be £178.9m, or 33% of the company’s shares. Raspberry Pi’s IPO, which was first rumoured back in 2021, is a reason to be cheerful about tech on UK public markets after the disappointment of Arm’s decision to list on #NASDAQ in 2023 and a series of recent take-private deals backed by US private equity. Encouragingly, CEO Eben Upton, said the quality of the interactions during the marketing process for the IPO has underlined their belief that London has the right calibre and sophistication of investor to support growing, ambitious technology businesses such as Raspberry Pi. Upton founded Raspberry Pi in 2012 to make computing more accessible to young people and it’s particularly pleasing to see the Raspberry Pi Foundation, the UK charity set up at the same time to promote computer science among young people, benefiting from the success of the business and its IPO. To date, the charity has received over £50m. Cambridge-based Raspberry Pi is about much more than just enthusiasts and education now though with its single board computers used worldwide and 72% deployed in industrial settings. In FY23, it reported revenues of $266m and a gross profit of $66m. I look forward to following Raspberry Pi’s journey as a listed company and hope its success might encourage other tech companies to follow suit. There are signs of an uptick in IPO activity more generally, with the likes of Chinese fashion giant SHEIN, Monzo Bank, BrewDog and Klarna also rumoured to be considering flotation in London. This post first appeared in TechMarketView's #UKHotViews. Be the first to read our views on tech news by signing up to the must-read daily email 👇 https://lnkd.in/euQSjG3d #LSE #uktech #STEM

6

Here's a blog post from me explaining what an IPO for Raspberry Pi Ltd would mean for the Raspberry Pi Foundation.

83

2 Comments

As a huge and life long Star Trek fan I find the future of Brain-Computer Interfaces (BCIs) a really interesting topic! 🧠💻 ✨ Nature Electronics named BCIs as the Technology of the Year in 2023, highlighting the sector's pivotal advancements. ✨ Synchron and Neuralink are leading the way with innovative trials, like Synchron's COMMAND trial and Neuralink's PRIME Study, aiming to enhance the lives of individuals with paralysis. ✨ Initiatives like the iBCI Collaborative Community are promoting interdisciplinary collaboration to overcome regulatory and clinical challenges. ✨ Research teams like NERV at Inria are developing algorithms that translate brain activity into commands, enabling thought-controlled interaction with the environment. 💡 Kudos to VC Newfund for their excellent work in their latest white paper on Brain-Computer Interfaces, shedding light on the investment landscape and future opportunities in this transformative field. #neurotechnology #innovation #braintech

5

Kudos to the UK government for the PSTI act for Consumer IoT. This is the summary: - Ban default passwords. Preinstalled passwords must be unique/device and sufficiently randomised. - Require products to have a vulnerability disclosure policy. - Require transparency about the length of time for which the product will receive important security updates. Hopefully more governments will follow. References: https://lnkd.in/ecTsed42 https://lnkd.in/eVUAMpNH #iot #security #defaultpolicy #legislation

4

Yesterday's post raised a question about combinational loops in ready/valid interfaces. In theory, any combinational path between ready and valid should be avoided. In practice, that can bloat a design with skid buffers, and misses out on some useful control logic (like that Synchronizer). Some combinational ready/valid paths are acceptable, with care. One day, I had to write this all down, so I had a consistent set of rules for ready/valid handshakes for my designs. It's worked well so far. https://lnkd.in/gv2tYgHV

30

7 Comments

Our latest performance update for Llama70B inference today achieves over 2,100 tokens per second, making it the fastest inference, period. And 16x faster than the fastest GPU solutions today. All made possible by the Cerebras Systems Wafer Scale Engine chip, that's 56x larger than traditional GPU chips. To try it out and learn more how this accelerate your #ml and #ai applications, go to https://lnkd.in/gYx4Y5XT

17

Here’s today's intriguing read One word: Neuromorphic What does it mean? It's the application of principles from how our brains work to create computing systems that mimic their structure, leading to more efficient computing https://lnkd.in/dkaUGYXT

2

Teleportation is here - on the quantum level.

4

1 Comment

A great CogX Festival 2024 yesterday at Royal Albert Hall talking about 4th industrial revolution that is being unleased by #artificialintelligence and #machinelearning technologies. Innovify has been for many years developing bespoke AI/ML applications for clients ranging from #startups to #enterprises. Whilst majority of the enterprises are currently looking to develop AI/ML capabilities, many are yet to lay the required groundwork. Here is a simple tool for you to check your organisation's readiness to start an AI/ML project. https://lnkd.in/d2bekp_3

15

If you are in the UK you maybe concerned with the constant rain at the moment. Under my #esg umbrella ☔️ (see what I did there? 😂) it is very interesting to learn that NVIDIA has teamed up with United Arab Emirates-backed G42 With new data showing that a ChatGPT search can use up as much as three bottles of water, AI companies are searching to make their models more efficient and better for the environment. Maybe we need this rain? NVIDIA and G42 are building a massive “climate tech lab” in Abu Dhabi, combining an advanced AI model with 100 petabytes (that’s 100 million gigabytes) of data. Looking at past weather data events, the model will be able to generate forecasts. It could also help us explore different approaches to fighting climate change: It’ll be able to simulate the global impact of different climate interventions, helping us make better decisions about which strategies to prioritise and what tools and models to use. One of my favourite quotes at the moment “when you create the ship you also create the ship wreck”. I am not for or against but we need to mindful and have a basic #AIliteracy https://lnkd.in/gKQRQSZg

10

3 Comments

Nvidia cut its teeth on GPUs, which Intel neglected, but AMD jumped in. (Intel wanted its CPU chips to do graphics as well so making GPUs redundant - and when GPUs were created - the gap was narrow. But then gaming became a thing - and the demand for advanced graphics was something that Intel did not anticipate.) And then they moved into AI chips. New startups in China has to start from way back - and then get their chip designs to SMIC which has limited bandwidth because they are creating cutting edge non-AI chips that China needs. Nvidia can leverage on their long and close working relationship with TSMC - while AI chip startups need to clamour for attention from SMIC. That means it is Huawei vs Nvidia - and that is goping as well as Intel vs Nvidia, "“The GPU software support ecosystem is much more entrenched around Nvidia and to a lesser degree AMD, and Huawei faces major challenges, both in producing sufficient quantities of advanced GPUs such as part of the Ascend 910C, and continuing to innovate and improve the performance of the hardware, given U.S. export controls that are limiting the ability of SMIC to produce advanced semiconductors,” Triolo said." In short, every one is waiting and hoping that Nvidia stumbles. Or TSMC.

4