Scott Stevenson
@scottastevenson
Followers
13K
Following
51K
Media
2K
Statuses
24K
Cursor for Transactional Lawyers: @SpellbookLegal Writing: https://t.co/5FBR8cexHv Metaphysical idealist
Joined November 2009
Meet Spellbook Associate: The first AI Agent for law. Two years ago we launched the first GenAI copilot for lawyers. We had no idea it would now be used by 2,500+ legal teams!. Today we're launching the first full-fledged agent that can break down legal projects into plans,
34
75
703
Warning to young technologists: there is a kind of person who is motivated to learn tech as a way to impress and gain cool points amongst other technologists. Maybe not knowingly, but subconsciously—they are socially motivated. They know Rust, they know Haskell, they only use.
394
741
10K
This is the first time I felt a game engine demo cross the uncanny valley
280
820
8K
The dark secret of capital gains tax during inflation—it taxes you for gaining nothing:. If you have $400k in retirement assets, and 25% more money is “printed”, your assets are now worth $500k, but you are no richer. The dollar just became 25% less valuable. Although you are no.
269
980
5K
This is one of the fundamental things people who struggle with creativity don't understand about it. The essence of creativity is "waste". So long as you are worried about "doing things efficiently and correctly the first time to reduce waste", you will not be creative.
"Leonardo da Vinci had so much unfinished work, such a shame.". No. You toss 80-90% of everything you create if you're a builder. If you don't have a workbook littered with "unfinished" projects you're not creating.
34
370
3K
Why sharpen your axe when the chainsaw is going to be invented tomorrow?
152
151
2K
If Ilya held his opinion that loosely, then who was the driving force at the board table?. It's looking increasingly likely that it was @adamdangelo , who was likely upset that his product launch got destroyed by Custom GPTs. "Not consistently candid" = "You didn't tell me you.
115
146
2K
A big trap in your 20's and early 30's is "vanity knowledge". Learning things that you think will impress people or get you to the next tier of status in your career. This is rampant in tech/software. A lot of the things "experts on the stage" talk about are fairly irrelevant to.
61
186
2K
Ilya is brilliant but he’s making a Founder 101 mistake by believing he can make a “straight shot” to the product he wants rather than iterating on useful products with customers. The problem with this mentality is that you are always levitating, never really touching the ground.
173
50
2K
As neural networks better at chess, they become more illegible: you can't explain why they make the moves they do in words. This will happen in the world at large as we approach singularity: you won't be able to explain in words why anyone does what they do.
90
166
2K
There is zero incentive for anyone to post about this defending the company. Anyone "playing politics" must defend the employee to earn charisma points. But that is deeply selfish and perpetuates illusions that ultimately hurt employees:. 1) She closed zero sales 4.5 months into.
I love that she recorded this. I love that people get to see what these giant corporations are like. From the human perspective, how could she ever improve with this feedback? I would be furious in her position. I've had to fire and lay off people, and I hope I did better.
178
104
2K
If we eliminate billionaires and give their money to everyone on earth, we each get $1285. $9T/7 billion people = 1285 dollars. That’s it. Im not saying billionaires should or shouldn’t exist, but a lot of people are suffering, and we need credible solutions that use Math.
98
149
1K
I predict Google/Bing search powered by LLMs will flop. Ask Jeeves & Wolfram Alpha also believed in this idealistic vision of "answering any question", and it was mainly a novelty. Most of us don't have weird trivia questions to ask every day. We use search like a phone book.
114
86
1K
@brianluidog A lot of things in life are like this, but the function is not to prevent someone who is sick from boarding, it’s to transfer liability to individuals who won’t be sued!.
2
3
1K
After working with LLMs for 2 years, I’ve noticed that humans have very different short-term context windows too. Humans with small-to-mid sized context windows develop deep “breakdown” strategies early in life to deal with virtually everything. Ways of breaking down complex.
60
121
1K
#1 thing I’d tell myself in my 20s: stop trying to maximize the “number of unique experiences” you have. A frantic collection of shallow experiences doesn’t add up to a deep experience. There is no “# of experiences” scoreboard. This is materialism in disguise.
30
91
1K
One of my biggest learnings as a founder and being around a lot of high performing people is: there are people doing work with the intensity that you would play Starcraft. Full days of incredible "actions per minute". I wouldn't have thought it possible unless I saw it.
31
42
1K
Step 1: Print a lot of money.Step 2: Asset and good prices skyrocket.Step 3: Tax as much of the "capital gain" as possible.Step 4: Blame Loblaws. 🇨🇦
36
189
1K
@nearcyan Standardization is good IMHO. It's also a warning sign: "Non-deterministic experiences lie here".
4
3
970
Today we launch our next leap forward in legal AI: Spellbook Reviews. Reviews adds precise redlines to your legal documents, live in Microsoft Word, based on your instructions. Chatting with AI was fun, but having AI actually do complex legal tasks before your eyes. opens up a
29
110
830
And if you introspect and find that you might be this person: Don't worry, pretty much everyone goes through the phase. The big question is whether you will do it for 20 years or outgrow it. If you outgrow it fast, you will be ahead of your cohort.
11
15
796
I am 36, and I can't think of a single moment in my life where making a "strategy document" paid off. What has worked for me is skipping to the end: Eg. just start prototyping ideas/copy/design extremely rapidly until finding something that feels like it's going to resonate with.
57
28
801
Companies saying that their goal is to make an LLM 10x more powerful than GPT4 are lost doing uncreative incremental innovation. IMHO the LLM project is *done* in the same way the SQL project was done in 2000. The interesting part of the next 10+ years is figuring out what we.
55
50
769
@catehall I think this is correct. We call “conscientious” kids gifted, because they perform well in a structured environment. But conscientiousness is a dark fucking trait to overcome in the real world. The things most worth doing are ambiguous and unstructured. Conscientious people are.
31
30
798
One reason startup success is so rare, is because the software engineering brain comes up with *terrible* idealistic ideas in the absence of customer feedback. We love clever architecture, modularity, extensibility and "platforms" that open up endless possibilities. All of that.
26
61
669
We're thrilled to announce that @SpellbookLegal has raised a $20m Series A financing, led by @inovia. In 2014 my musical instrument company got a legal bill that nearly zeroed our bank account. In 2019 we launched Rally to help lawyers draft more efficiently using document
65
54
639
Absolutely mind blowing. GPT knows enough about computers and the internet that you can boot up an imaginary computer, browse an imaginary internet, and make API requests to imaginary servers. And it all works perfectly.
14
101
675
@soychotic If you think a $20b company hasn’t had a team empirically analyze the ROI on this, you might be the one missing an elephant in the room!. Marketing is often about displaying the intoxicating platonic ideal of a product. A bottle of Coke is not that interesting or good. The.
167
7
610
Using GPT-3 to explain and rewrite sections of a contract based on your instructions!
16
81
650
One of my biggest mental shifts since using GPT3, is that I now see every organization as having an API. Every email as an API call. Everything is already an API, it just took us awhile to teach computers the protocol.
18
57
617
Being a generalist is often just a symptom of having high agency and driving outcomes. If you’re a high agency teen and you *really* want to make a video game, you will learn how to program, make graphics, maybe music, writing, and a bunch of other stuff. You aren’t really.
19
66
611
@paulg So much so that we end up keeping those technologies around just to marvel at them: Eg. mechanical watches vs. Digital watches, live classical music vs. recordings.
9
13
543
Deeply believe that this is the future of creativity, not Suno. Creatives mostly want electric bicycles, not to board an AI train that can’t be granularly controlled.
🎵💿Built an MCP that lets Claude talk directly to Ableton. Now you can create music with just prompts!. Here’s a demo of me creating a lush, 80s synthwave track in just two prompts. It picks the right instruments, creates melodies, and adds effects like reverb and distortion 🔊
47
44
572
How my brain has started to evaluate candidate "seniority" after hundreds of job interviews. This is a vaccine against "Manager Mode", which can promote a lot of Level 1 people to the top of an organization.
11
30
524
The idea that we should strive to maximize our experiences in life so that we have no regrets on the deathbed is anti-buddhist and anti-mindfulness and the root of a lot of anxiety. There is no scoreboard, there are no chips to cash in at the end of life. The idea that we are.
23
38
501
In our seed financing round, we booked 80 meetings with VCs using cold emails.
Marc Andreessen on why VCs ignore cold emails — and why that’s not unfair. “The way the top-end venture capitalist firms work is they’ll basically take you seriously if you come in introduced by somebody they’ve worked with before, and they won’t take you seriously if you don’t.”
26
25
474
Finding alpha by intentionally making bad opening moves: . Magnus Carlson sometimes intentionally makes "bad" opening moves in order to push the board state into less charted territory, where he can outmanoeuvre competitors. This is a technique you can use in business and even.
11
20
455
@thesamparr Two reasons to be wary of index funds:. 1. It’s *way* easier to beat the market with a small amount of capital. Moving around billions is insanely difficult because your buys and sells move the market and everyone is trying to frontrun you. I am a hobby investor, and my active.
33
17
451
Generative AI for legal is just getting started 👀
15
38
401
I had non-stop sinusitis for about 10 years. Tried absolutely everything to fix it. Neither me nor Dr. could figure it out. Accidentally stopped eating gluten a couple months ago and the problem went away completely. Life is full of these weird “lock in keyhole” problems. Try.
56
12
439
I love Arc and I hope they succeed, but. Every software company that has tried this kind of hip & cool marketing has ultimately failed. Reminds me so much of Ello, the "cool" Facebook (. Ello's unique character gave them instant buzz and they felt.
51
15
423
Everything rests on whether you can find a couple hours a day to actually move things forward vs. being stuck treading water, maintaining status quo. There is an infinite difference between standing still and moving forward. Whether you find those 2 hours or not is a night and
3
27
411
@ClaraSorrenti It can seem like capitalism is making things worse when most of us have only lived fairly short lives and don’t have much to compare too, but when you look at the data, we’re being absurdly and unexpectedly successful at eradicating poverty:
15
36
370
It is very easy to win charisma points by not pointing out difficult truths. Many wealthy people get there by making everyone around them feel good and avoiding difficult truths. I believe this is an extreme hidden form of selfishness that hurts a lot of people, and that we have.
18
16
374
It's been cool to watch OpenAI do product-market fit search at such massive scale. You can imagine how this innovation would have been crushed at most companies:. "We're an infrastructure company like AWS not a consumer app company! What the heck is this ChatGPT thing? We need to.
10
17
361
@david_perell Choosing to play a stoic defense does not necessarily dictate how you play offense.
13
20
357
Vibe coding personal software for anything you need is wild. There is going to be a Github for non-devs to share and branch their creations, and it's going to be massive.
45
26
353
@ID_AA_Carmack @elonmusk Imagine a new “historical” view that could display the evolution of a topic or idea over decades on Twitter! Whole new kind of value. Twitter has barely scratched the surface in unearthing historical value.
5
9
330
Low agency critics who don’t build things love saying that every “chat with PDF company is now dead” due to ChatGPT launching PDF uploads. They need to justify their lack of action to themselves: “aha! I knew I was right to sit on my butt and not build anything on top of GPT!”
Wow, within a day every single of the over 400,000 users of ChatWithPDF just vanished. Straight line down to zero. All b2b contracts cancelled, my GPU providers also decided to cut access to my embedding model hosting last night.
17
20
330
@FuckCarsReddit They're actually really energy-efficient. Only need to heat above freezing. Whether suburbs are good is another debate, but no one is going to generate value for themselves, their family or for society by shovelling. We have a moral duty to not waste our time on stuff like that.
7
4
317
A path to high agency:.1. Person steps out of a highly structured environment to an unstructured environment (eg. Is left alone to write a PhD thesis, starts a company, etc).2. Person has gut-churning anxiety for ~6 months as they realize they don’t know how to self-motivate.3….
8
17
333
o3 described the danger of systems thinking in product development better than I could:
19
25
343
People don’t think AI music will possibly be appreciated by the masses—because who is going to emotionally connect with AI?. But this connection has already been faked in decades of pop music. People thought Daft Punk actually played synthesizers and drum machines rather than.
92
34
321
A common pitfall in product development is what I call “tech tree development”. The first half of my career was spent thinking this way. This comes from the engineer’s desire to believe they are building a glorious castle in the sky, with many foundational and reusable.
15
24
332
Why is Deepseek so mimetically fit? . I have never seen AI news discussed more quickly by such a broad variety of people.
64
7
334
Spellbook's AI copilot for lawyers is now available to law students & faculty, completely free. We've been blown away by the number of educators reaching out for access for their classes--and also by how many schools are contemplating banning AI from the classroom. Read more 👇
14
57
290
"Alone in Newfoundland", a Studio Ghibli movie made up by AI. (More below!)
3
46
291
Devin for Legal is coming. This is the first sneak peak of Spellbook's agent. It reviews a VC term sheet and starts editing a set of financing documents. It's not perfect yet. But it's improving quickly. Agents are about to leap from their GPT2 era to their ChatGPT era.
15
44
291
When in crunch mode, my task management system becomes a plain text file, and I get 5x more done that way. Is a text file actually the best task system? Are all other forms of task management simply a form of procrastination and clinging to structure?.
61
17
295
@adamdangelo @tszzl Imagine the blowback Adam had to face from his employees: . "Dude aren't you on the board of OpenAI? Why did we spend the last 6 months building this? They screwed us.".
4
13
283
“Are GPT3-based apps defensible?” is similar to asking “Are SQL database apps defensible?”.
19
15
268
I do not see AGI or ASI as concrete things, they are abstract ideas, but their concrete forms will be numerous. It will not be a magic spell you can simply cast on anything. We have invented fire, now we need steam engines, power plants, candles, and many more things.
8
9
273
@adamdangelo Poe's product launch, with the team probably worked for months on, had no visible impact on search interest. A brutal feeling for any founder
3
6
261
It’s amazing what happens when you change your mindset from: “I need to repair everything that’s decaying” to “I need to create value faster than things decay”. If your house is decaying at 2% per year and you create 10% a year in new value, you are way ahead. The great trap for.
15
29
263
1/ We're thrilled to announce that @SpellbookLegal has raised $10.9M USD!. Last summer, we launched the first generative AI copilot for lawyers. Since then we've had our heads down shipping 30+ features, expanding to 600 legal teams & growing a waitlist of 54,000!. 👇
55
34
252
@pmddomingos It’s not a secret, they blatantly acknowledge it, while competitors waste time trying to build better LLMs. They know plug-ins + autogpt will be the next big thing and they’re leading the way there.
4
6
248
@Lola_lmao7 I think it’s comprehended, it’s just that they intuitively feel that there is *a lot* more uninhabited space—which looks true:
4
2
216
Maybe you will for individuals. But for those that need to be competitive to survive, like businesses and investment funds, illegibility will become the new norm. Legible stories will become the sign of a losing entity stuck in the past. While winners will look like RenTech.
4
7
227
Everyone in startups should try to get good at poker, in order to see the absurd compounding power of acting on imperfect information. Repeatedly acting on low sample size data is one of the most powerful things startups do to outmaneuver incumbents. Eg. We surveyed 5 lawyers to.
6
15
228
@Suhail Were you holistically knowledgeable when you had success with Mixpanel? . The pattern I’ve seen is that the naive hackers with Beginner’s Mind frequently win vs. the knowledge accumulators. Don’t put too much pressure on yourself. Being “behind” could be an advantage.
9
8
224
Maintaining mental self-reliance can be lonely and gruelling. The world that lies beneath social narratives is so alien that you will, on some days, experience immense pressure to conform and question your own sanity. Here are some readings for those days ⬇️.
4
25
223
The ability to stop thinking in stories, IMO, will be one of the most powerful skills in the next 10-20 years.
8
15
203
AI is moving so fast that the VC model of capital deployment may be too slow. Startups themselves may need to become the “edge compute” for capital deployment. When new opportunities arise overnight, companies will use the incredible leverage of AI to capture them almost.
17
20
213
@adamdangelo @tszzl This feels like a psychopathic tit for tat: "If you destroy what I am working on, I will destroy what you are working on". Again, speculation, but I don't feel guilty about it given the circumstances.
5
5
206
All the "picks and shovels" LLM tooling for eval and observability is not solving the core problem for app developers:. LLM prompt chain evaluation is so subjective that it crosses over into the realm of UX design, more than engineering. Many of our successes resulted from 80.
23
21
213
@_mattneary The average workplace post-ZIRP is a disaster IMHO. Years of your board being wrong, your CEO being wrong, your boss being wrong--because no one was held to real constraints. A fantasy version of reality spiraled out of control, where everyone was just mimicking eachother without
8
20
206
@sama 100%. It took me almost 10 years to realize how I was wasting my life on fake prerequisites. Their appeal is that they are structured and unambiguous. There is no apparent risk in reading another book or tutorial, you have high confidence that you will finish and get a sense of.
5
19
198
Big news: Spellbook now integrates with content from Thomson Reuters Practical Law. Often lawyers don't want pure generative content. They want trusted precedents as starting points. For two years, customers have asked if we would work with the TR team to build a bridge between
8
24
188
This transition isn't hypothetical. We are deep in the midst of it now. It's why the world is becoming increasingly confusing, and stories of meaning are harder to find. Any plausible story of meaning (eg. fighting climate change) gets latched onto for dear life.
7
6
193
Humans are anxiously grasping for concrete story beats to hold onto while the world increasingly can only be understood as a matrix of numbers.
2
8
178
The trickiest traps are the ones that make you think you’re escaping another trap. I thought I was escaping the shallowness of materialism by focusing on collecting “experiences”, but it was just another kind of grasping.
3
8
180
@waitbutwhy Counterpoint: kinda weird how Apple can’t figure out how to design their own database software and all these other companies with way fewer resources managed to pull it off. Maybe the LLM part is too much of a commodity and they are trying to figure out deep UI level innovation.
13
6
189
There are three ways we could deal with this:.1. Simulation: simulate simpler legible environments - we do this now with video games.2. Go Backwards: intentionally cause ourselves to reverse progress.3. Intuition: stop thinking in words.
4
4
177
After years of running super lean, working remotely and renting in various co-working spaces, @SpellbookLegal has its first office that's fully our own
33
2
185
@achtera lol you are confusing this with a wealth tax. The vast majority of those that live in Canada/US pay capital gains tax at some point.
3
3
172
In real life, when you “skate where the puck is going to be”, you look crazy. In real life the hard part of this is not the technique itself, but being comfortable looking nuts. This deeply goes against our tribal instinct. It is inevitable that agents will work for coding and be.
i’ll defend devin and their team here. most of the critique here are things you’d hear from a senior engineer (dinosaur) who’s watching over your shoulder and constantly negging you. it’s completely regarded . most of his arguments are strawman .- devin used an incorrect version
21
21
178
O3 is the first time I've really felt like I was delegating to team member who is smarter than myself, and who has pretty good taste.
8
6
179
ChatGPT has nothing to do with search. It's a productivity tool, a writing tool, a creativity tool. It has no place in a phone book! It belongs in the places where creators work.
12
18
175
@soychotic For all the nice people replying: yes I think corporations are always right and never ever make mistakes and all CEOs are perfect angels. Actually what I’m saying is: there’s no way this is not a staged rollout where metrics aren’t being carefully monitored. If it’s in every.
11
2
160