From: aidotengineer
Bar Yaron, a partner at Amplify, focuses their investments on data and AI companies, specifically targeting technical founders [00:00:25]. This focus highlights the significant area of venture capital interest in the evolving fields of data and artificial intelligence.
Influential Figures and Trends in AI
During a survey conducted among 100 AI Engineers, several influential figures in the AI research space were identified [00:07:07]:
- Jeff Hinton [00:06:21]
- Yan LeCun [00:07:22]
- Andrew Ng [00:09:09]
- Fei-Fei Li [00:09:13]
- Yoshua Bengio [00:09:19]
Sam Altman, while influential, was not among the top eight most influential AI researchers according to the survey [00:07:56].
Considerations for Choosing AI Models
When AI engineers choose an AI model, the top consideration, surprisingly to some, is cost [00:10:51]. Other important factors include:
- Latency [00:11:08]
- Accuracy/Performance [00:11:28]
- Open-source versus Closed-source options [00:12:49]
- Capabilities [00:13:18]
Industry Buzzwords and Challenges
Some buzzwords that AI engineers are reportedly tired of hearing include “AGI” (Artificial General Intelligence) [00:15:38], “RAG” (Retrieval Augmented Generation) [00:16:25], and “Prompt Engineering” [00:16:25].
Hot Takes on AI Development
Industry experts offered several “hot takes” on the future of AI and its development:
- At least one of the major entities currently training large AI models may no longer be doing so by the end of the year [00:01:55].
- AI could serve as effective therapists [00:02:13].
- A significant trend will see the majority of deployed AI models running on-device within a year and a half [00:03:34]. This points towards smaller, more specialized models orchestrated together, reducing reliance on larger models and external data processing [00:03:40]. This could impact cost-effective AI strategies and IT infrastructure management.
- AI could facilitate future dating apps, with AIs dating each other [00:04:13].
- Transformers may not be the final AI architecture, with the potential for models to be built using biological materials [00:04:19].
- Future AI conversations are predicted to involve multiple bots in a room rather than just one-on-one interactions [00:04:40].
Jobs at Risk and Industry Benefits
When asked about jobs most at risk of AI disruption, survey respondents identified:
- Data entry [00:20:43]
- Content creation/writing [00:20:14]
- Software engineers [00:17:44]
The industry expected to benefit most from AI is healthcare [00:21:00], indicating potential for significant efficiency improvements with AI.
Key AI Papers and Tools
The most influential AI paper in history, according to the surveyed engineers, is “Attention Is All You Need” [00:17:49]. A popular AI tool among engineers is Cursor [00:18:04].
The State of AI Engineering Survey
A broader survey on the State of AI Engineering is being conducted, aiming to provide transparency within the industry regarding tools and workflows used by AI Engineers [00:21:39].