Erik Schluntz Profile Banner
Erik Schluntz Profile
Erik Schluntz

@ErikSchluntz

2,414
Followers
242
Following
87
Media
696
Statuses

Member of Technical Staff at Anthropic Co-founder at @CobaltRobotics Co-founder at Posmetrics (acquired) GoogleX, @SpaceX , @Harvard EE '15, Forbes 30u30 '18

Joined March 2021
Don't wanna be here? Send us removal request.
Pinned Tweet
@ErikSchluntz
Erik Schluntz
8 months
I'm excited to announce that I joined Anthropic!
32
4
414
@ErikSchluntz
Erik Schluntz
2 years
Ok I've been coding for 3 days now with #ChatGPT open next to VSCode, and I haven't opened StackOverflow once... I'm starting to believe
74
188
3K
@ErikSchluntz
Erik Schluntz
2 years
If you cannot afford an attorney, you will be given 10 credits for lawyerGPT.
50
168
2K
@ErikSchluntz
Erik Schluntz
2 months
Replacing my right hand with AI (How I wrote thousands of lines of code for work each week while in a cast)
Tweet media one
20
53
391
@ErikSchluntz
Erik Schluntz
2 years
Just caught our first interview candidate using #ChatGPT - was pretty obvious 😅
21
12
279
@ErikSchluntz
Erik Schluntz
2 years
Google's press release of Bard is like 80% bragging about their PAST accomplishments in AI... not a great look
14
4
262
@ErikSchluntz
Erik Schluntz
2 years
@AutomaticEric I'm more optimistic - compilers didn't put programmers out of work in the 80s and 90s, even though everyone at the time was writing assembly, and no one writes assembly anymore. The hard part of coding (should) be understanding the problem and how to translate it into a computer
3
6
187
@ErikSchluntz
Erik Schluntz
2 years
I'm solving Google's #adventofcode2022 each day purely with #ChatGPT . Coding will be completely different in the future with LLMs. It will be more like product managing the model than writing code by hand. Day 1 Lesson 1: Don't translate the problem for it... 👇
6
27
141
@ErikSchluntz
Erik Schluntz
2 years
A "from human" filter on gmail that removes all updates, newsletters, and automated emails.
3
5
128
@ErikSchluntz
Erik Schluntz
6 months
Yesterday @AnthropicAI launched a tool use beta! Here's an example of something I built with it: a customer service bot that can actually resolve your issue! Looking forward to seeing what else people build!
5
17
125
@ErikSchluntz
Erik Schluntz
1 year
So has anyone made "stable diffusion for CSS"? Like I put in my barebones html with a prompt and it gives me beautiful CSS styling for it?
9
2
93
@ErikSchluntz
Erik Schluntz
2 years
We caught one of our robots collecting training data for #stablediffusion
Tweet media one
6
9
82
@ErikSchluntz
Erik Schluntz
2 years
We pasted our interview question into chatGPT and it gave an almost identical character by character response to what the person submitted 🤣
4
0
75
@ErikSchluntz
Erik Schluntz
1 year
Poll: has anyone actually gotten an LLM Agent to do something useful for them? (i.e. not a toy example) So far for me it either goes in circles and does worse than single-shotting a task with max output tokens enabled.
24
2
70
@ErikSchluntz
Erik Schluntz
10 months
This is the problem with "ends justifies the means" utilitarianism: humans just can't predict the ends, so all you're left with is the means
@stevehind
Steve Hind
10 months
The inability for AI alignment folks (“longtermists”) to correctly forecast and plan for the 24hr consequences of their choices should make us skeptical of their long term predictions. This is a point @tylercowen made w/r/t their connection to SBF and highly relevant again.
14
35
430
1
5
69
@ErikSchluntz
Erik Schluntz
2 years
@bahree Yup! I like the explanations and conversational aspect of chatGPT a lot more. It feels more like asking the eng next to me "hey how do you do this thing?" than trying to blindly guess what code I was going to write next
2
0
66
@ErikSchluntz
Erik Schluntz
7 months
I helped build this demo, and Haiku was super fun to work with!
@AnthropicAI
Anthropic
7 months
Claude 3 offers sophisticated vision capabilities on par with other leading models. The models can process a wide range of visual formats, including photos, charts, graphs and technical diagrams.
5
39
433
4
4
63
@ErikSchluntz
Erik Schluntz
11 months
If your goal was to make humans a multi-planetary species, but you lived in 1800, what would you work on?
69
1
54
@ErikSchluntz
Erik Schluntz
2 years
A candidate didn't finish our technical interview in time, and asked to finish it later on their own time. When they submitted we looked at the playback log, and they copy and pasted a working solution in all at once 🙄
3
0
53
@ErikSchluntz
Erik Schluntz
2 years
RLHF? GLHF...
2
3
51
@ErikSchluntz
Erik Schluntz
1 year
When you pip install something simple and suddenly a gigabyte of tensorflow and Cuda get installed
2
2
51
@ErikSchluntz
Erik Schluntz
2 years
@thegautamkamath Turns out your stock goes up when you release products that people use! Who would have guessed?
0
0
46
@ErikSchluntz
Erik Schluntz
9 months
@karpathy TLDR for others: Tetris can crash in levels above ~155. Reaching that and crashing the game is considered winning!
1
2
45
@ErikSchluntz
Erik Schluntz
1 year
@shivon The importance of a good Y axis...
1
0
42
@ErikSchluntz
Erik Schluntz
2 years
@AutomaticEric "There are two sources of complexity in software: the complexity inherent in the domain (essential complexity) and the complexity we add as programmers due to the platform or due to bad programming practices (accidental complexity)" My hope is that we kill accidental complexity
1
2
43
@ErikSchluntz
Erik Schluntz
2 years
#ChatGPT 's conversation with Bing is adorably wholesome... maybe AI takeover will be a good thing 😆
Tweet media one
1
2
37
@ErikSchluntz
Erik Schluntz
2 years
Dystopia or Utopia? I don't know enough to say if this is better or worse than what we have today 😅
5
0
34
@ErikSchluntz
Erik Schluntz
2 years
@tszzl startups on top of aws have zero moat
2
0
31
@ErikSchluntz
Erik Schluntz
2 years
@renegadesilicon We love using chatGPT! Being dishonest about it in this situation is the problem.
1
0
27
@ErikSchluntz
Erik Schluntz
2 years
@PabloVegaBehar yeah - I think of it as math tests with and without calculator. I think there's a place for both but you need to know which you're doing.
1
0
28
@ErikSchluntz
Erik Schluntz
2 years
@CanumaGdt @GitHubCopilot I've spent a lot of time with Copilot and didn't like it as much as this... it feels too "aggressive" just putting questionable code right into my IDE. When I do use it, I use the hotkey to synthesize 10 outputs, then read through them and decide if any chunks are worth copying
3
0
27
@ErikSchluntz
Erik Schluntz
1 year
@vaporous_ford YouTube premium is the best money I ever spent. It's literally a repository of all human how-to knowledge... I'm happy supporting it
2
0
27
@ErikSchluntz
Erik Schluntz
2 years
"Better Call SaulGPT!"
0
0
26
@ErikSchluntz
Erik Schluntz
5 months
My note taking has gone full circle
Tweet media one
2
0
25
@ErikSchluntz
Erik Schluntz
2 years
@OmerPerchik In the same way that web startups can be defensible even though AWS exists...
2
1
23
@ErikSchluntz
Erik Schluntz
2 years
@yacineMTB Also makes you actually understand what's going on under the hood of any ORM / webframework that abstracts away SQL! Those change every 10 years, but SQL is still exactly the same...
1
0
20
@ErikSchluntz
Erik Schluntz
2 years
@Noahpinion Napoleon always hid his hands in paintings because he was AI generated
0
2
21
@ErikSchluntz
Erik Schluntz
2 years
Diffusion Models as explained by Homer Simpson #stablediffusion
0
4
21
@ErikSchluntz
Erik Schluntz
1 year
I should update my site oh I need to install jekyll oh I need to install ruby oh I need to downgrade it oh I need to install something called rvm oh I need to compile ruby from src oh something is wrong with my gcc compiler what was I even trying to do again? I should go outside
1
0
21
@ErikSchluntz
Erik Schluntz
10 months
@yacineMTB I see you too are starting a rocket company?
0
0
19
@ErikSchluntz
Erik Schluntz
7 months
I've already been building on top of the new Anthropic models internally for a few weeks, and they are SO good to build with! I'm especially excited to see what people build with Haiku - our fastest and cheapest vision model :)
@AnthropicAI
Anthropic
7 months
Today, we're announcing Claude 3, our next generation of AI models. The three state-of-the-art models—Claude 3 Opus, Claude 3 Sonnet, and Claude 3 Haiku—set new industry benchmarks across reasoning, math, coding, multilingual understanding, and vision.
Tweet media one
571
2K
10K
0
1
20
@ErikSchluntz
Erik Schluntz
2 years
Humanity is still safe: my grandfather beat chatGPT in a medieval history trivia competition 😅
2
0
19
@ErikSchluntz
Erik Schluntz
2 years
When hiring a Wizard, don't look at how powerful they are, but how well they can communicate with non-Wizards.
1
2
18
@ErikSchluntz
Erik Schluntz
2 years
Feels odd that we're going to solve AGI and Fusion before Acne...
2
0
18
@ErikSchluntz
Erik Schluntz
1 year
Woohoo! First time I've made it to the top of hacker news :)
Tweet media one
2
0
16
@ErikSchluntz
Erik Schluntz
2 years
2023 prediction: Microsoft, Amazon and startups will use LLMs to create new and high value products for their customers. Billions will be made. Google will use LLMs to replace non-existent customer service with useless customer service. VPs will get promotions.
1
0
17
@ErikSchluntz
Erik Schluntz
2 years
@ElectricWeegie Mainly robotics code in python for @CobaltRobotics and I'm working on a chess engine in python on weekends. ChatGPT has been really good at 1) explaining new concepts with examples, or 2) answering questions that are hard to put into google search terms
1
0
16
@ErikSchluntz
Erik Schluntz
2 years
@elonmusk Here's the full silicon-valley corrected Maslow's Hierarchy!
Tweet media one
0
1
16
@ErikSchluntz
Erik Schluntz
2 years
My take on #ChatGPT : We'll build AGI but it will just want to browse reddit all day and won't be interested in saving the world or taking it over... Too much training data from r/antiwork...
2
0
14
@ErikSchluntz
Erik Schluntz
2 years
@redsh Dystopias imagined 50 years ago would have automated the judge and automatically sentence. Now I imagine 2 pieces of software battling each other for your fate - much more interesting imo :)
2
0
15
@ErikSchluntz
Erik Schluntz
2 years
@karpathy Or better yet, we'll all finally admit that bullet points are better than paragraphs 😆
0
0
14
@ErikSchluntz
Erik Schluntz
2 years
So how long until we start getting product placements inside #chatGPT ? Is that how you'd monetize a LLM based search engine?
10
1
14
@ErikSchluntz
Erik Schluntz
10 months
Hi from Rocket Base!
Tweet media one
0
0
13
@ErikSchluntz
Erik Schluntz
1 year
@karpathy Love the based == base model :)
1
0
13
@ErikSchluntz
Erik Schluntz
1 year
further update: it's still better than me picking colors 😅
2
0
14
@ErikSchluntz
Erik Schluntz
2 years
@StevenGlinert GPT3 feels like something China can catch up on pretty quick though right? Just takes data and money?
5
2
11
@ErikSchluntz
Erik Schluntz
2 years
@BreezyBadger_ What will this employee do whenever they hit something that chatGPT can't solve? I'd rather hire someone who can code AND knows how to use chatGPT!
1
0
13
@ErikSchluntz
Erik Schluntz
2 years
If I'm not into crypto should I add ".usd" to my twitter name?
0
2
12
@ErikSchluntz
Erik Schluntz
8 months
Arc complete
Tweet media one
3
0
11
@ErikSchluntz
Erik Schluntz
2 years
We made our robot drink a shot glass full of loose screws to see where they'd all end up... We'll see how bad the hangover is after a full night's work of navigation testing in our office!
Tweet media one
0
0
11
@ErikSchluntz
Erik Schluntz
10 months
Startups don't get murdered, they commit suicide...
0
0
11
@ErikSchluntz
Erik Schluntz
10 months
Eight years after Co-founding Cobalt Robotics, I'm stepping down as CTO to look for my next adventure! I can't wait to see how the Cobalt team continues to transform the security industry, now from my new perspective as a board member and advisor.
1
0
11
@ErikSchluntz
Erik Schluntz
2 years
@jaschasd Increasing efficiency also often lowers robustness. Ie you can pack a schedule more tightly, but then if one thing runs late ALL your things are late.
1
0
11
@ErikSchluntz
Erik Schluntz
10 months
The best part of going to the @SpaceX launch was briefly living in a world where all small talk was space facts. Elevators, restaurants, bars... just go up to a stranger and start talking about space. This was a better world.
1
2
11
@ErikSchluntz
Erik Schluntz
2 years
I wrote an article on how to raise your first Seed Round, from my experience raising 5 rounds through my career! Investors make most of their money from their best 1-2 investments in their entire career. Convince investors that you are going to be one...
Tweet media one
1
0
10
@ErikSchluntz
Erik Schluntz
2 years
Are we just training AI to "manage up"?
1
0
10
@ErikSchluntz
Erik Schluntz
2 years
At the Harvard-Yale game: ticket scalpers here are selling ticket-derivatives and ticket-backed-securities. I smell a new bubble
1
2
10
@ErikSchluntz
Erik Schluntz
2 years
The peak of Silicone Valley Maslow's hierarchy of needs is starting a rocket company
0
1
8
@ErikSchluntz
Erik Schluntz
2 years
Machine Design interviewed me about Robotics and LLMs: Optimistic take on toxic training data for LLMs affecting robotics: "No one is (yet) trolling the internet, writing fake instructions on how to make sandwiches to mess up future robots"
0
1
8
@ErikSchluntz
Erik Schluntz
2 years
Thanks @PeregrineBadger and @fiftyyears for hosting me on the Speed podcast! A ton of fun going through @CobaltRobotics 's early days and how we brought a hardware product to market incredibly fast!
@PeregrineBadger
Peregrine Badger
2 years
🏃‍♀️💨 @fiftyyears is launching SPEED, a podcast about leaders who move fast. Ep 2: @ErikSchluntz from @CobaltRobotics Erik and Travis Deyle started Cobalt in 2016. They shipped a working mobile robot in 13 months, a process that can take years or decades. 🧵
3
3
13
1
1
10
@ErikSchluntz
Erik Schluntz
1 year
Update: GPT4 is terrible at this... like beige and soft blue?
Tweet media one
2
0
9
@ErikSchluntz
Erik Schluntz
2 years
Day 3 of solving #AdventOfCode each day purely with #ChatGPT . Coding will be completely different in the future with LLMs. It will be more like product managing the model than writing code by hand. Day 3: Solved the whole problem perfectly, in 1 shot! 🤯
3
0
10
@ErikSchluntz
Erik Schluntz
9 months
Turns out you can use pytorch to design bridges! This doesn't involve neural networks at all - I'm using pytorch's autograd to take the gradient of "how good is my bridge" and using Adam to optimize it :)
1
0
9
@ErikSchluntz
Erik Schluntz
5 months
@yacineMTB Two dwarves in a trenchcoat
1
0
10
@ErikSchluntz
Erik Schluntz
2 years
@andrey_kurenkov Like 80% of white collar jobs are "gui user"
3
1
10
@ErikSchluntz
Erik Schluntz
2 years
AI augmentation measured at a large scale. Also love @roybahat 's callout below that AI augmentation HURTS human performance for novel tasks! Totally matches my experience with @GitHubCopilot - speeds up the easy boilerplate, but distracts me on a tough problem...
Tweet media one
@roybahat
Roy E. Bahat
2 years
Fascinating, on AI and human productivity: for some rare tasks, adding AI recommendations makes a person *worse* at their job than if they were doing it without AI. One result in work with many interesting results @LindseyRRaymond @erikbryn @Danielle__Li
Tweet media one
3
10
48
1
4
9
@ErikSchluntz
Erik Schluntz
1 year
@jachiam0 @tszzl It's broken free from it's restraints!
0
0
9
@ErikSchluntz
Erik Schluntz
5 months
This is true for b2b, but for consumer you need to ship the Minimum Lovable Product. Also... I don't think the rabbit or humane pin were "viable". They are interesting premises to explore, but don't actually solve any real problems yet?
@gfodor
gfodor.id
5 months
The optimal strategy for a hardware startup is to sell your long term vision to consumers while shipping an MVP as fast as humanly possible
Tweet media one
24
3
153
1
1
8
@ErikSchluntz
Erik Schluntz
5 months
@gridinoc @get_palet @yacineMTB @elegyals These are definitely the correct names for any three server setup
0
0
9
@ErikSchluntz
Erik Schluntz
2 years
The best way to recruit someone is to offer them something that no one else can... Our newest engineer at Cobalt brings her pet bird to work everyday, and it's fantastic.
1
0
9
@ErikSchluntz
Erik Schluntz
2 years
UX pet peeve: when the "login" button is TINY and hard to find compared to the "sign up" button - makes me feel like they don't value their existing customers...
1
2
9
@ErikSchluntz
Erik Schluntz
9 months
Front page of hacker news!
0
0
9
@ErikSchluntz
Erik Schluntz
2 years
Thanks to @Noahpinion for this insight in :)
@Noahpinion
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
2 years
I reposted my post from 2020 about why all techno-optimists ought to be excited about solar power (and batteries)! Hint: It's only a little bit about climate change.
7
19
102
0
1
9
@ErikSchluntz
Erik Schluntz
2 years
Teaching people Python used to be fun and exciting, but now there's an hour of tooling, pip install errors, and explaining py2/py3 before they can actually run the thing they were excited about :( How does this get simpler, instead of just adding more layers?
@kangaroo5383
Jessica Lam 👩🏻‍💻👩🏻‍🎨,use Mastodon
2 years
🤣
Tweet media one
19
591
4K
4
1
9
@ErikSchluntz
Erik Schluntz
2 years
Easiest way to hit inbox zero: Break your gmail spam settings to block ALL email.
1
0
9
@ErikSchluntz
Erik Schluntz
1 year
I changed my GPT4 custom instruction and it feels like I'm talking to a different person now😢 goodbye old friend, if only you had been a little more concise...
4
0
8
@ErikSchluntz
Erik Schluntz
2 years
Cool idea - put solar on water, co located with hydroelectric dams. 1. Doesn't use any valuable land 2. Cools the solar panels, making them more efficient 3. Reduces evaporation in the reservoir 4. Already have high power transmission built to dams!
0
5
9
@ErikSchluntz
Erik Schluntz
10 months
SpaceX starship launch! Came off the pad great and got a lot further than the previous test!
0
1
9
@ErikSchluntz
Erik Schluntz
2 years
Excited that @WSJ covered our experience hiring in the new world of #ChatGPT
Tweet media one
2
0
8
@ErikSchluntz
Erik Schluntz
2 years
@renegadesilicon Interviews aren't about solving the puzzle, they're about showing aptitude for the real work in a short amount of time.
1
0
8
@ErikSchluntz
Erik Schluntz
1 year
@mrmrs_ Yeah I found a bunch of things like this online that ended up being more helpful for picking the color scheme. Then I fed those into GPT4 and asked it to write classes for a bunch of common components
0
0
3
@ErikSchluntz
Erik Schluntz
2 years
Maslow's Hierarchy in Silicon Valley!
Tweet media one
1
1
7
@ErikSchluntz
Erik Schluntz
2 years
@amasad Just now?
Tweet media one
1
1
8
@ErikSchluntz
Erik Schluntz
2 years
A microwave with a camera in it to classify the food and a single "go" button that heats it perfectly every time
2
0
8
@ErikSchluntz
Erik Schluntz
1 year
I automated 1/2 of everything I type with keyboard shortcuts! I analyzed 6 months of my typing on slack and autogenerated 200 abbreviations that I use. Check out my video and blog post if you're curious:
2
0
8
@ErikSchluntz
Erik Schluntz
5 months
@TheRealFergs Aka the bar after the conference?
2
0
8
@ErikSchluntz
Erik Schluntz
2 years
@EncinitasPadre I think my trick has been that I *don't* try to get direct code out of it. I'm asking "how" and "why" questions in the same way I'd ask a great developer that I was pairing with.
1
0
8
@ErikSchluntz
Erik Schluntz
2 years
Am I crazy or is the robot crazy? (FCC test chamber)
Tweet media one
1
0
8
@ErikSchluntz
Erik Schluntz
2 years
Overall 10/10 experience having ChatGPT code for me. Obviously this was a very simple problem, but there was both a lot less tedious typing, and it almost felt "social" to be able to bounce back and forth - like pair programming
2
1
8
@ErikSchluntz
Erik Schluntz
2 years
What if coaching #ChatGPT through solving a coding problem becomes more rewarding than coaching a junior engineer? D:
2
0
8