Carolynn Levy, Inventor of the SAFE | Hacker News Reader

Carolynn Levy, Inventor of the SAFE

meridian.mercury.com

112 pointsbaxtr24 comments

I appreciate the conscious effort to be concise:

“Writing a short legal document is actually quite hard. It’s a cliche — the simpler it is, the harder it is to do, and us lawyers are so used to writing for each other and not thinking about how short things should be.” — Carolynn Levy

“The idea itself wasn't the innovation. The concept was similar to the convertible note. The innovation was writing such a short document without any of the complex pieces.” — Kirsty Nathoo

In general, I find it crazy that we allow legal documents to be so complex. And not just complex, but comped to the point of being less clear than something stated in simple, plain language.

Legal matters, both civil and criminal, are so of the most important things in a person’s life. How do we allow language that is only decipherable by “experts”? Even if you understand the language, the “correct” interpretation of that language is governed by decades and centuries of inscrutable case law. To me, this is a tragedy. Laws and contracts should be made in simple language understood by a majority of the populace. People have a right the understand the rules that govern them.

In my experience there are three sources of complexity in contracts:

(1) inherent complexity in the subject of the contract

(2) a complex form being applied to a simple subject

(3) bad drafting

The first is sometimes unavoidable. For example, if you're drafting a loan document that has an interest rate based on a benchmark, it's probably not enough to reference that benchmark. You also want to include some provisions that will specify what happens if the benchmark ceases to be published. This kind of fulsome treatment makes contracts longer and more complex.

The second is a a result of the way most lawyers work. It's far quicker (and therefore cheaper for the client, at least in the short term) to take an existing form and make changes to adapt it to a new situation than it is to create a bespoke contract. This reuse of forms propagates archaic language, needless "boilerplate" clauses, and the like. Many commonly encountered provisions are akin to junk DNA. They're included when the template is replicated but long ago ceased to have any function. Example: counterparts clauses or clauses "allowing" electronic signatures. [1]

To me, the solution to this issue is to build contracts differently. Instead of starting with a recent precedent document and choosing "save as," we need to assemble them by carefully choosing well-drafted clauses from a library--think of a recipe for a convertible note that has "#include preemptive-rights" or something like that.

Finally, drafting contracts well is a very specific skill. It's not a case of using the right "magic words." Rather, it's being able to describe the parties' bargain clearly and precisely. This isn't taught in most law schools, and most lawyers never bother to learn it systematically. Again, they're rushing to adapt existing forms that a colleague has blessed, and very shy about creating something new.

I think most practitioners are aware of these factors but feel they can't really change them due to being caught in an hourly billing treadmill. In the article, Carolyn Levy says she couldn't' really tackle this problem until she went in house. That has been true for me as well.

[1] https://www.adamsdrafting.com/counterparts-an-example-of-cat...

But what I think you disregard is the practitioner’s preference to copy and paste arises because either that’s specific language that has either been litigated and a court has ruled on what those words mean (to that court anyway), or you’re trying to implement the same deal as all the other contracts with the same provision, so why write it differently?

In the best case scenario, the new, carefully crafted language will be interpreted exactly the same way as the old language, and in the worst case scenario, it will be interpreted as meaning something totally different - because if the drafter meant to say the same thing as everyone else, why didn’t they just use the words everyone else uses?!

I’m not particularly defending this state of affairs, but precedent counts for a lot - not just in common law legal decisions, but in contracting as well.

Really interesting. I did not know the origins. I do know the Levys (they used to have the early classes over for BBQs and it was awesome) and one thing I didn't pay attention to until now is how I would just always send crazy ideas their way and they would humor them and explain the legal issues around them. Now it all makes sense—Clevy was breaking new ground in the legal world (if there's one thing I never paid any attention to until recently its the legal world). Thanks for putting this history together Mercury!

I’m lawyer for private companies, though very rarely for VC-backed companies, and to my paranoid mind, the SAFE is at the end of the day not really much more than a term sheet on a handful of key economic terms. Which, in a high trust environment, is just fine the vast majority of the time and the reduction in transaction costs is almost always worth the trade-off! For instance, no one reads the deal documents they’ve paid hundreds of thousands of dollars for (if not more) after they’re signed … until there’s a commercial disagreement, and then every word, and sometimes punctuation marks, can make or break a case, so the docs get read VERY closely at that point. (This is an exaggeration, but directionally true.)

But statements like the following from TFA (which I enjoyed quite a bit and found quite interesting) seem like they obviate the protections that are the sole reason for equity investment documents to exist:

It seems to be that, after a certain point, the utility and success of the SAFE really is contingent on a founder's worldview — how much effort they put into tracking dilution, how closely they read the document when they sign off on terms, and whether they keep their promises to investors, even when they don't have to.

If you don’t have to keep it, it’s not a promise! If the fundraising system relies on norms and reputations to enforce baseline expectations, you frankly don’t need contracts - a handshake, an email memorializing key terms and the whisper network is enough to keep founders and investors in line. But for any kind of reasonable legally enforceable rights and remedies, that’s gotta be documented - and unfortunately, usually at great length and expense.

(Side note: private equity sponsors frequently, though not as often as they used to, push back on their counterparties in negotiations by telling them “trust us, we’re in the reputation business - if we screw you, no target/management team/issuer will ever take our money again.” When the sponsor has to choose between its IRR and its reputation, guess which it picks?)

I think the SAFE is a cool instrument. I think it’s great for start-ups existing in the right cultural setting. But it’s less a legal innovation than a thoughtful short-form summary of a standardized set of commercial terms.

Question to any VC lawyers who happen to see this - have there been many SAFEs litigated?

> For instance, no one reads the deal documents they’ve paid hundreds of thousands of dollars for (if not more) after they’re signed

The financing docs don’t costs hundreds of kilobucks, more like an outrageous 20K or so (with the investors’ lawyers’ bill coming out of proceeds — scandalous, as that’s what the 2% management fee is for!). An acquisition can cost a lot more.

And actually I do always read the deal book. Because they are huge piles of cut/paste & patch, they are inevitably full of inconsistencies, overlooked deviations from the TS etc. None of which matters to get the deal signed, but as you say: potentially fatal down the road if a situation comes up when you actually need to consult it.

Also notes were never a big deal to put in place; I used them on a deal just before COVID and it simplified a lot of things. SAFES are handy too, but I didn’t like the change to post-money which I find much less founder friendly.

> The financing docs don’t costs hundreds of kilobucks, more like an outrageous 20K or so (with the investors’ lawyers’ bill coming out of proceeds — scandalous, as that’s what the 2% management fee is for!). An acquisition can cost a lot more.

One distinction I should have mentioned - the vast majority of my deals use LLCs to raise capital. That increases the number of points that need to be negotiated by probably two orders of magnitude compared to an investment in a Delaware corp. Way more efficient to rely on the corporate code, fiduciary duties and non-waivable duties of care than to re-invent virtually every component of economics, governance and control on every deal. Yet here I am…

But of course, another part of the reason the bill is so high is because we’re not just copying and pasting disparate provisions - we’re making sure the doc matches up with the commercial deal. But yeah, you could half-ass an LLC agreement in an afternoon, and it’d be good enough 90% of the time probably.

It's still expensive to raise Series A. Like why isn't there a very standard document that can significantly reduce the cost of financing

There are. Always ask for counsel to use NVCA docs.

It's expensive because partner time for this kind of transaction is $1200-2000/hr and their juniors will be a fair %age of that rate too.

You shouldn't need a partner for this kind of deal. In fact, startups should never have to pay for partner time and I never have at a company with less than about $50MM in revenue (even with firms like Gunder, Cooley, Wilmer, MoFo etc).

for some debt financing structures the legal cost can be up to $500K:

Launching a Financial Product: How to Choose the Right Funding Structure

https://a16z.com/2022/10/10/launching-a-financial-product-ho...

Well, for some specialized bankruptcy debt according to that page. But I have done venture debt (e.g. from someone like WTI or SVB) and never had legal bills as high as they talk about there. Then again they talk about $75MM in debt; the most debt I've taken on was about $25MM and it has always been cheaper than any equity financing.

In general I don't like to get creative on this stuff -- I leave that to the actual execution of the business.

Do you have the pre-money version? I wanted to reads the difference before I even saw your comment. I imagine it might be archived somewhere, but it's no longer on the main site.

And why are the change less founder friendly?

So I'm not a lawyer, and I haven't litigated anything, and this is not legal advice, but have raised capital (ventures debt, safe's, and Series-A) and seen how PE companies work/do acquisitions. In thinking about these documents, it's worth remembering that for these companies there are basically only two outcomes: 1. Do a priced round (Series-A) that converts the investment. 2. Or, go bankrupt/run out of money and stop operating.

For #1, conversion i.e. Series-A, tend to be done by a small number of legal firms that everyone knows, and VC's have their own legal council too. While there can be plenty of negotiating on terms; I've never heard about conversion being contentious. Compared to many other things, these are 'happy events.' Both sides need to 'opt-in' to the agreement. And the law-firms and VCs have strong incentives to play close to the book.

For #2, bankruptcy's or wind down, there's usually nothing to fight over. The firm has no money, and no-one has put value to the assets.

This is extremely different from many corporate transactions where a) the firm may have a lot of tangible/valuable assets, b) the seller is wealthy/sophisticated/has ulterior motives and c) the buyer is wealthy/sophisticated/has ulterior motives. So there's tons of potential motivation, and reason, to fight over things.

So I think both the finite ways SAFE's are used, and the finite outcomes/incentives to the counterparties significantly reduce the 'reasons' for litigation - compared to the vast majority of corporate agreements/contracts/legal proceedings.

There is a third outcome which is a problem for SAFEs: the company has moderate-but-not-amazing growth and never raises an equity round. The founders are able to just keep growing it, paying themselves good salaries and eventually dividending out profits to themselves. Because SAFEs have no maturity date and have no rights to dividends, the investors get nothing and are basically stuck.

Such "zombie" companies are effectively a write-off for venture investors given the power law dynamic of their funds. Usually the SAFEs will be negotiated away (modest payout or written down to zero) as the 10-12 year lifetime of the fund draws to a close.

Most parties are good about coming to an amicable conclusion since a bitter dispute would irreparably harm both parties' reputation in future endeavors.

irreparably is overused

on the contrary, almost nobody remembers anything where they weren't a main character

I'm not talking about zombies. These are profitable, growing companies. Just not venture scale.

Often you can work something out with founders, but the fact that the SAFE doesn't have anything to say about these situations is a bug.

SAFEs share in dividends: "However, if the Company pays a dividend on outstanding shares of Common Stock (that is not payable in shares of Common Stock) while this Safe is outstanding, the Company will pay the Dividend Amount to the Investor at the same time."

Of course, most startups will never declare dividends.

I immediately assumed this article was about the Scaled Agile Framework (SAFe) and was pleasantly surprised.

Yea, I thought so as well and was immediately terrified about their safety.

I already planned how I would balance hate speech and censorship when writing a response about how Carolynn Levy should be considered public enemy number one of developers and the murderer of real agile development.

It’s funny because it’s almost the opposite. “Simple Agreement for Future Equity”, an intentionally abbreviated legal document to make deals easier and, dare I say it, more agile. As opposed to SAFe, a methodology written by consultants, for consultants.

Reply on news.ycombinator.com