Google DeepMind's AlphaGeometry and AlphaProof, have shown abilities in solving complex math problems that near the very best human minds. AlphaGeometry excels in geometry, reaching a level close to a human International Mathematical Olympiad (IMO) gold medallist, while AlphaProof has achieved the same level as a silver medallist in the IMO competition.
AlphaGeometry uses a neuro-symbolic approach, combining a neural language model with a symbolic deduction engine. The language model identifies patterns and relationships in geometric data and proposes useful constructions, such as points, lines, and circles. The symbolic engine uses formal logic to make deductions and find solutions based on these suggestions. This approach allows it to tackle geometry problems requiring multiple steps. Importantly, the solutions generated are verifiable and use classical geometry rules.
AlphaProof is a reinforcement-learning system that focuses on formal math reasoning. It can translate natural language problem statements into formal statements. The system generates solution candidates and proves or disproves them using a search over possible proof steps in the formal language Lean. In a virtuous circle, the proofs are then used to reinforce the language model, enabling it to solve increasingly challenging problems.
A key aspect of these systems is their ability to train on large datasets of synthetic data, generated without human demonstrations. AlphaGeometry used a "symbolic deduction and traceback" process to create 100 million unique examples of varying difficulty.
AlphaProof trained by proving or disproving millions of problems. This approach overcomes the data bottleneck that has limited the use of formal languages in machine learning. In the 2024 IMO competition, a combined system of AlphaProof and AlphaGeometry solved four out of six problems, earning a silver medal equivalent score. This achievement marks the first time an AI system has reached this level of performance in the IMO.
Online education company Chegg has found itself with the honour of being the first company to at least ‘publicly’ admit that Generative AI appears to have had an adverse impact on its business model. On its most recent earnings call on May 1st, Chegg’s CEO Dan Rosensweig said that “since March we saw a significant spike in student interest in ChatGPT” which they believed was having an impact on new customer growth rates. The result was a 48% drop in the share price on the day to levels where it has remained since.
Meanwhile IBM a company which is no stranger to developing AI technologies having pioneered Watson over a decade ago and being involved in AI powered technology that was the first to beat a reigning world Chess Champion - also weighed in on the impact of AI. In an interview with Bloomberg, IBM’s CEO Arvind Krishna suggested that they will slow and even suspend hiring in job functions that they believe could be done using today’s AI technologies.
This they believe would impact around 3% of its 260,000 strong workforce (roughly 8,000 positions) over the next five years. These roles are predominantly in back office, not customer facing positions such as human resources, and accounts.
There have been other high-profile examples of AI appearing to have a meaningful impact on jobs. After announcing that it was going to use AI to help generate material to enhance its content and quizzes in early 2023 just weeks after the media company, Buzzfeed announced a 10% reduction in its workforce. More recently the company admitted that it had used generative AI to write around 40 of the travel guides it had used online recently. It is safe to say this is just the tip of the iceberg with AI undeniably automating tasks in the service industry just like robotics has in the manufacturing industry.
Recent filings have revealed that Zipline, one of many drone delivery startups globally, recently managed to secure $330m in a series-F funding round at a valuation of $4.2bn. Since its initial venture round 12 years ago the company has now raised over $820m according to Crunchbase making it one of the best funded
Zipline started out in the developing world flying critical medical supplies such as blood and vaccines in Rwanda and other nations because healthcare is more readily funded and regulatory approvals for unmanned flights come faster. However, it has also been growing its commercial delivery business, with partners including Walmart, GNC, Pagliacci Pizza and Associated Couriers. They target one million autonomous deliveries by the end of 2023 after passing the 600,000 level in April.
Source: Zipline, YouTube
Zipline’s newest drone the P2 utilizes a two-stage delivery system, when the drone arrives at the target destination it then lowers what they call the delivery droid down on a zipline while the main part of the drone continues to hover above to enable ultra precise delivery.
Both the main drone and the smaller delivery droid are autonomous – the delivery droid (pictured above) utilizes a series of fans to keep it stable and allows it to gently land deliveries with ultra-high precision.
Another one of their innovations includes charging stations which look like lamp posts where the drone can autonomously land and recharge while it is waiting for its next delivery.