A.I. Executive Briefing #2

The A.I. Executive Briefing is an expert weekly curation of A.I. news by our research team, shared externally now because we feel there’s too much hype & noise in the market. The same content will be distributed through?this substack.

NEWS ROUND-UP

1. 400+ Chat GPT Plug-ins

2. UK gets access to Foundational Models

3. EU AI Act passes parliament

4. The Evolution of LLMs

5. BCG, Mickinsey & Github A.I. reports

VENTURE FUNDING

6. France’s?Mistral AI?raises a?$113M seed round?to take on OpenAI

7. Unbabel Acquires Bablic in an undisclosed round

8. AI video creation startup Synesthia raises $90m in a Series C

9. Real Estate AI startup EliseAI Raises a $35 Million Series C

10. Legal AI company EvenUp raises $50 million?at $325M valuation

11. Salesforce pledges to invest $500M?in generative AI startups

12. ML startup SiMa.ai Raises $13M?in Funding

News Round-up

1. ChatGPT plugins hit 400

The total number of ChatGPT plugins released apparently crossed the 400 mark this week. When we checked on June 11, 100 plugins were added in preceding 24 hours alone:

However, while many people think ChatGPT will be the next digital interface and their plug-in ecosystem will become the next Apple app store, below is a screenshot of the actual ChatGPT plugin store when we access it through our pro account. It’s unclear who builds these plugins, how you build them, how they’re approved, and this UI is every product designers worst nightmare — there are 56 pages of plugins you have to click through with no categorization or way to explore these plugins.

2. OpenAI, DeepMind and Anthropic to give UK early access to foundational models for AI safety research

Following his announcement that the UK will host the first major global summit on AI safety later this year, Sunak reiterated the UK's position as a leading democratic AI power, with top AI labs, including DeepMind. An expert task force on AI foundation models, backed by ￡100 million in government funding, will lead on cutting-edge AI safety in the UK.

3. EU AI Act passes parliament

The EU has completed their draft of what A.I. regulation would look like within Europe. It passed their parliament making them the first government body to put pen to paper and publish concrete ideas. The AI Act is not law yet and is pending negotiators at EU institutions, such as the EU executive body and 27 member states to also pass the act. Mainly the idea is that all A.I. systems would need to undergo risk assessments before putting the tech into everyday use, akin to the drug approval process. Additionally, there are concerns over how exactly the specific EU states will enforce the laws in the A.I. Act specifically against marginalized groups or those deemed enemies of the respective state. With that, here are the takeaways:

1. Unacceptable risk posed by A.I. systems that are banned outright

• Real time facial recognition & Biometric tracking for any social scoring purposes.
• Cognitive behavioral manipulation of people or specific vulnerable groups

2. High Risk use cases of A.I. systems will have to be registered in an EU database:

• Biometric identification and categorization of natural persons
• Management and operation of critical infrastructure (Food, water & Electricity)
• Access to and enjoyment of essential private services and public services and benefits

3. Limited risk use cases of A.I. pertain mainly to generative A.I. systems

• Publishing summaries of copyrighted data used for training

4. The Evolution of LLMs

Given all of the developments in the LLM space and the variety of use cases companies are touting for their special models, I offer you two charts. The first a very detailed timeline of the development of Open source and Closed source models from Nathan Lambert , who leads the RLHF Team at HuggingFace. The second is a more high level overview of the primary transformer models being developed by the current A.I. leaders in the space put together by Alex Irina Sandu

5. Boston Consulting Group, Github, & McKinsey all recently published reports on A.I. development use cases and views. Highlights below:

• BCG - While 80% of business leaders said they were using A.I. regularly, only 20% of frontline employees said the same. This upside Pareto Principle lends itself nicely to the A.I. bubble we see forming. 80% of leaders are using A.I. but only 20% of their workforce is actually using it to add value to their businesses. If A.I. is meant to automate low-level tasks and only a fifth of the frontline workers are using it, that is a clear disconnect from the fervor seen at the top levels of organizations.
• Github - 92% of U.S.-based developers are already using AI coding tools both in and outside of work. However, developers still say DevOps is the most time-consuming thing they’re doing at work besides writing code is waiting on builds and tests.
• McKinsey - Current generative AI has the potential to automate work that absorb 60% - 70% of employees’ time today. Additionally, Generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually. Here is a solid chart of the use cases McKinsey identified across four major sectors:

Funding News

6. France’s?Mistral AI?blows in with a?$113M seed round?at a $260M valuation to take on OpenAI

This company is pre-product and was founded some point in May. Funding like this round is indicative of a bubble similar to one we saw emerging in 2019-2021 with seed rounds inflating in size and valuation to no end (Remember Clubhouse? Neither do we). According Antoine Moyroud from Lightspeed, this round was justified on the fact that the Mistral team was able to consider the “second order effects of LLMs.” Regardless of their $100m insights, an emerging profitable narrative for entrepreneurs in this cycle seem to be to raise in relation to OpenAI’s valuation, claiming sine they are competing with them their company should be worth X percentage of their valuation (Link)

7. Unbabel Acquires Bablic

AI-powered Language Operations (LangOps) platform, acquired Bablic, a Tel Aviv, Israel-based provider of a cloud-based solution supporting all localization and translation management requirements of websites and web applications. As LLMs continue to conquer and learn human languages, there will be smoothing effect of how people can communicate with each other. (Thinking of Naval’s Airchat). In the knowledge economy, ease of communication will be the forefront currency.

8. AI video creation startup Synesthia raises $90m in a Series C to become a unicorn

Synthesia is an AI video creation tool that creates videos from plain text. Users can make polished video presentations in-browser with no professional expertise required from just text. Text to video start-ups open up the potential to democratize content creation. Lots of content creation barrier flattening with the text-to-XYZ trend we are seeing in the generative A.I. space

9. Real Estate AI startup EliseAI Raises $35 Million in Series C Funding

EliseAI is a real estate company that developed an “AI leasing” platform to automate billing, notices and tenant communications. They seem to be expanding into general enterprise business use case as well as healthcare. A clear departure from their expertise, hopefully there is some transferability in their back end systems that make this move viable.

10. Legal AI company EvenUp raises $50 million?at $325 million valuation

EvenUp provides a product to personal injury law firms that automates the workflow for demand letters, using generative AI and a proprietary legal dataset. Its customers have seen a 30% increase in payouts, in addition to time savings from drafting legal claims, CEO Rami Karabibar told Reuters.

11. Salesforce pledges to invest $500M?in generative AI startups

Salesforce doubled their fund size from their announcement earlier this year. They have also stated that they are putting A.I. at the forefront of their User Experience. A lot of companies from Accenture to PWC have been advertising new A.I. funds and workforces. Perhaps it will contribute to ballooning inflations as public corp VCs attempt to save their stock price by “getting involved” in the AI cycle.

12. ML startup SiMa.ai Raises $13M?in Funding

The following is taken straight from the press release: “Led by Founder and CEO Krishna Rangasayee, SiMa.ai is a Machine Learning company delivering a software-centric, purpose-built MLSoC platform” —> can anyone tell us what they do??? If anyone can ELI5 what this company does you will get a shout out next week!

Reader Questions

Send us a message with any questions/comments/thoughts on anything A.I. related and we’ll try to answer them in our next release.