Remember those saxophonists that thread their way through packed nightclub dancefloors, enthusiastically duetting to whatever bass-boosted banger the DJ is playing? For reasons unknown, AWS Summit 2023 decided that was exactly what we needed at their vast registration hall at 8AM. Accompanying the saxophonist and festival-style speakers was a drummer, presumably for good measure.
Cloud computing isn’t glamorous, and perhaps that explains the musical intervention. But what it lacks in glamour, cloud computing – and AWS in particular – makes up for in infrastructural importance. 83% of unicorns globally are powered by AWS, and it has accompanied many of them throughout their rapid scaling. London’s 2023 AWS Summit was a chance to learn what the world’s most adopted platform is doing to assist the success of its customers. Hint: it’s gone big on generative AI.
AWS Summit 2023
‘It’s easy to innovate when times are good. Less so when they’re not.’ This was the core message from Tanuja Randery, AWS’s MD for the EMEA region in her introduction to the keynote speeches. She pitched AWS’ innovations as not just as a next-gen platform with the tools to take your business to new heights – but as a trusted partner to help optimise costs when times are rough. Randery was equally keen to point out AWS’ investments into the UK, and their bullish estimate that cloud computing can add more than £400B to the UK economy over the next few years.
Leading the keynote was Swami Sivasubramanian, AWS Vice President of Data and Machine Learning who went all in on AI. Keen for AWS not to be seen as a big corp trying to keep up with the kids, he emphasised Amazon’s time-honoured commitment to these technologies, noting their e-commerce recommendations engine is driven by machine learning, as are the paths that optimise robotic picking routes in their fulfilment centres. This ain’t AWS first rodeo.
Swami wasted no time telling us that generative AI, powered by large ML models (Foundation Models (FMs)) are revolutionising the creation of content and ideas across various domains. The central tenet of his talk was that what sets FMs apart is their ability to be customised for specific business needs, offering a unique customer experience. Although generative AI has caught the attention of customers, Swami told listeners there’s still much untapped potential.
AWS has recognised the demand for easy access to FMs and has therefore introduced Amazon Bedrock: a service that provides access to FMs from AI21 Labs, Anthropic, Stability AI, and Amazon via an API. In essence, Amazon Bedrock addresses key customer concerns by offering high-performing FMs, seamless integration into applications, and data protection. Swami claims it democratises access to generative AI by enabling builders to easily find, customise, and deploy FMs using familiar AWS tools without managing infrastructure. Customers can choose from cutting-edge FMs like the Jurassic-2 multilingual LLMs, Anthropic's Claude, and Stability AI's Stable Diffusion for text-to-image generation.
Accompanying his explanation was an example of an imagined customer looking for help with their idea for an urban walking shoe. By leveraging the aforementioned FMs, he showed us how to take a kernel of an idea to a fully fleshed product description (using Claude), visualise it (using Stable Diffusion), get a load of social copy (using Jurassic-2) and then finish up with the keywords (using Titan Text). Given the evident technical prowess, however, they could have chosen a more engaging example: the shoe looked ugly, my eyes hurt reading the social copy and it sounded like a bland product idea - but what do I know?
Both Compare the Market and Arup also got on stage to share how they’ve been using AI in their work, thanks to AWS. Will Cavendish, Arup’s Global Digital Services Leader, nicely illustrated that by combining geospatial and machine learning technologies using Amazon SageMaker JumpStart, Arup was able to help tackle the urban heat island crisis. This was done by going granular on the detail in the data, and subsequently realising nature-based solutions.
AWS Startups
I also had the chance to speak with Kellen O’Connor - General Manager of the AWS business for Startups in EMEA. He tells me that AWS is no stranger to startups – in fact it was these customers they focused on in the beginning. He cites Marc Andreessen’s comment that the cost of founding a startup dropped by a factor of 100 when AWS was launched, so great was its ability to offer affordable and scalable compute resources. He points out that being so intimately connected to the startup ecosystem allows them to build products straight off the back of customer feedback. Amazon Aurora – a flagship offering – is one such example, only developed thanks to AWS listening to complaints about legacy vendor database licensing models. Until last year it was the fastest-growing product in AWS history.
AWS Startups does a variety of things – with the goal of ‘helping companies scale to become the next Wise, Monzo, Babylon or Skyscanner’ as Kellen simply puts it. Those are all UK unicorns and longstanding AWS customers. But our focus today is on the accelerators and what makes them stand out.
The accelerators are also built from startup feedback, although not as directly as other products. Kellen says they got into the accelerator game after understanding current offerings weren’t always right – not every founder wanted to give away equity or follow a set curriculum. Targeting early-stage startups, AWS gives credits and technical help to a range of promising businesses. They aren’t playing around: one program, AWS Activate, has distributed over $2B in credits to over 100,000 companies.
AWS has accelerators covering everything from healthtech, defence and fintech in Africa: UK-wise there’s a B2B SaaS of which this year’s cohort has just been announced. Most notable are the Gen AI and Female Founders in AI outfits, which are global. Kellen tells me two UK startups - Theia Insights and Flawless AI - have made the cut for the Gen AI cohort, whilst the Female Founders in AI cohort will be announced shortly. He also pointed out only 20% of the current AI workforce is women.
AWS may be hopping on the hype train, but their offering is certainly one of the most sophisticated and customer-focused - as demonstrated by Arup - and they also did well to ground their internal commitment to AI.
