Analyzing AngelList Job Postings, Part 2: Salary, Equity Benchmarks, and AI-written Text Detection

Analyzing AngelList Job Postings, Part 2: Salary, Equity Benchmarks, and AI-written Text Detection

In the fast-paced world of tech startups, one crucial aspect of building a successful company is hiring the right talent. And when it comes to hiring, salary and equity play a significant role. In this article, we will delve into the benchmarks for salary and equity in engineering jobs in Silicon Valley, as well as introduce a new AI classifier that can detect AI-written text.

When it comes to offering equity to early employees, founders need to strike a balance between generosity and strategic allocation. The danger of being too generous, especially with the first few hires, is that you may end up giving away more equity than necessary. This equity could be better utilized to strengthen future job offers, attract more investments from investors, or retain decision-making power for founders. It is important to note that the benchmarks we present here are based on a medium-sized sample and should not be considered hard-and-fast rules.

For employees 2 through 13, the salary range increases for higher paying jobs. The 20th percentile salary range is $75k - $100k, the 50th percentile salary range is $85k - $125k, and the 80th percentile salary range is $100k - $150k. These figures provide a general idea of the salary expectations for engineering jobs in Silicon Valley.

Equity allocation also follows a pattern based on the order of hiring. For the first hire, it is recommended to allocate 2% - 3% of equity. Hires 2 through 5 typically receive 1% - 2% equity, while hires 6 and 7 receive 0.5% - 1%. As the number of hires increases, the equity allocation decreases. Hires 8 through 14 receive 0.4% - 0.8% equity, hires 15 through 19 receive 0.3% - 0.7%, and hires 21 through 27 receive 0.25% - 0.6% equity. Finally, hires 28 through 34 receive 0.25% - 0.5% equity. It is important to remember that these figures can vary depending on the specific circumstances of each startup.

A useful way to frame these numbers is by adding "up to" before each range. For example, a typical 6th hire may receive up to 0.5% - 1% equity. Similarly, designers among the first four hires can expect up to 1% - 2% equity, occasionally only 0.5%. Designers among the next five hires may receive up to 0.5% - 1.0% equity, and for employees 10-30, the equity allocation ranges from 0.2% - 0.5%.

It is crucial for both founders and potential employees to have realistic expectations during the hiring process. Unrealistically high expectations from an employee can be a waste of time for both parties involved, leading to missed opportunities for the candidate and a failed interview process for the company. On the other hand, founders should be cautious if an employee's expectations are too low. Exploiting this may result in long-term resentment and hinder employee retention.

Now, let's shift gears and discuss an exciting development in the field of artificial intelligence. Researchers have recently trained a classifier capable of distinguishing between text written by humans and text generated by AI. While it may not be possible to detect all AI-written text reliably, this classifier can provide valuable insights and help mitigate false claims of human-written text.

In evaluations conducted on a challenge set of English texts, the classifier correctly identified 26% of AI-written text as "likely AI-written" (true positives). However, it also falsely labeled human-written text as AI-written 9% of the time (false positives). It is important to note that this classifier should not be used as a primary decision-making tool, but rather as a complement to other methods of determining the source of a piece of text.

The classifier does have some limitations. It is highly unreliable when applied to short texts below 1,000 characters, and even with longer texts, there is a possibility of incorrect labeling. Additionally, the classifier performs significantly worse when applied to languages other than English and is not reliable for code detection.

The classifier itself is a language model that has been fine-tuned on a dataset consisting of pairs of human-written text and AI-written text on the same topic. The dataset was collected from various sources believed to be written by humans, including pretraining data and human demonstrations on prompts submitted to InstructGPT.

In conclusion, understanding the benchmarks for salary and equity in the tech industry can help founders make informed decisions when hiring early employees. It is essential to strike a balance between generosity and strategic allocation of equity. Additionally, the introduction of AI classifiers that can detect AI-written text opens up new possibilities for identifying the source of written content. However, it is important to recognize the limitations of such classifiers and use them as a complement to other methods.

To summarize, here are three actionable pieces of advice:

1. Founders should carefully consider the equity allocation for early hires, keeping in mind the long-term impact on fundraising, job offers, and decision-making power.
2. Potential employees should have realistic expectations during the hiring process, avoiding both overly high and low expectations.
3. When utilizing AI classifiers for text detection, it is crucial to understand their limitations and use them as a complement to other methods of determining the source of a piece of text.

By incorporating these insights and taking proactive steps, founders and potential employees can navigate the hiring process more effectively and make informed decisions for the growth and success of their companies.