Samuli Laato (@LaatoSamuli) header
Samuli Laato (@LaatoSamuli) profile picture

Assistant professor of information systems at @uniTurku

Joined January 2020
Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)Photo by Samuli Laato (@LaatoSamuli)
Samuli Laato retweeted
Sufficiently advanced agentic coding is essentially machine learning: the engineer sets up the optimization goal as well as some constraints on the search space (the spec and its tests), then an optimization process (coding agents) iterates until the goal is reached. The result is a blackbox model (the generated codebase): an artifact that performs the task, that you deploy without ever inspecting its internal logic, just as we ignore individual weights in a neural network. This implies that all classic issues encountered in ML will soon become problems for agentic coding: overfitting to the spec, Clever Hans shortcuts that don't generalize outside the tests, data leakage, concept drift, etc. I would also ask: what will be the Keras of agentic coding? What will be the optimal set of high-level abstractions that allow humans to steer codebase 'training' with minimal cognitive overhead?
208
517
127
4,321
477,768
Samuli Laato retweeted
I don't read/write code these days, I only instruct an LRM. But it's not because I think LRM code quality is perfect, or even good. And it's not because I think my instructions are always being perfectly acted upon -- I know they aren't. The thing is: LRMs don't just write your code, they unlock new ways to control what you're doing and maintain understanding of your system, which are about as effective as reading code. You can better test your codebase, you can audit specific components, you can spin up visualizations, you can ask the LRM to red team, etc. You can do all of these things at high speed -- you couldn't before. The benefits of reading/writing code can now be achieved in new ways. All in all, the ROI of writing code by hand is no longer looking good, not because LRMs got perfect but because they're so fast at doing things that you can develop new workflows around them, workflows that end up being more productive than the old workflows.
124
154
37
1,808
172,303
Samuli Laato retweeted
Dear Astra, When I ask you to fix something, I mean write the three lines that fix it. Not 400 lines of code, 17 tests, and a 12-page essay documenting the bug’s final days. Best, Hegar
533
427
88
12,149
623,231
.
128
1,308
126
16,113
475,753
There are two openings I hate: 1. e4 e5 And 1. e4 c5 2. Nf3 d6 3. Bg5 Both last moves cause involuntary urges to quit chess for good.
26
Palworld is so decidedly ahead of recent Pokemon games in every aspect that now Scarlett and Violet seem like the cheap knock offs.
1
40
Samuli Laato retweeted
16 years ago today, Blizzard announced that StarCraft II had already sold more than 3 MILLION copies. It had only been out for about a month. There was a time when a new Blizzard PC game launching felt like an event the entire gaming world stopped for.
40
61
9
694
23,937
Samuli Laato retweeted
We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of the Lord of the Rings, a 1M token budget (~$10) and asked for three js render of it. Opus went off for ~2 hours and wrote 5500 lines of code that (procedurally) rendered the story. It's kind of janky but fun. But it's a bit mindboggling that the LLM has to place and orchestrate various polygon assets in (x,y,z) coordinates and write code that animates it all, and that it even does anything at all. I also like this kind of examples because no one in their right mind would ever spend the time to write something this custom but LLMs have all the stamina and patience in the world, so it's an example where we go from "no one would ever do this" to "sure, why not, it's ~free". There might be a lot more. But I'm excited about creating hyper custom worlds that you can imagine dropping players into, e.g. here to participate in the LoTR story as a spectator NPC, or one of the characters, or etc. Something like an ephemeral GTA of X on demand. Last thought is that the domain of worlds/games exposes a weakness in LLMs: they can't easily audit their work because they aren't able to efficiently and natively perceive videos or play games within them. Here, Opus 5 had to very slowly and painstakingly take screenshots at different points, and it messed up a few times and created a bunch of jank. An example of raw capability (multimodal, gameplay) that I think is still quite lacking.
1,804
2,312
753
29,785
6,322,117
Samuli Laato retweeted
Programming is not about code, just like music is not about notation. It is the art & science of managing complexity through layers of abstraction. AI is simply a part of it.
127
384
47
3,159
124,556
Samuli Laato retweeted
The technology Hurricanes fans are using to watch their team play in Vegas 👀 (via @Canes)
134
658
68
10,683
1,117,116
AI slop on YouTube has gotten so bad that I get excited when I find a 3+ year old video.
2
65
Samuli Laato retweeted
Replying to @VictorTaelin
Another thing: what you get from writing things yourself isn't just the code. It's an improved understanding of what the code does. That mental model is what lets you come up with further improvements, or invent a different way of doing things. You can't come up with ways to improve a blackbox you don't understand. For most projects this doesn't really matter, because the code is the only thing you need. But if you're doing something novel, if you're doing research, the code is not the most important part. Understanding what the code does is the most important part.
15
44
7
666
29,995
For me personally genAI tools have only increased cognitive load. Everything in my work happens faster, and it's increasingly more demanding to distinguish gold from slop.
2
3
215
Samuli Laato retweeted
Hot take: If you don't even want to read the paper once, maybe don't put your name on it.
The funniest thing about #arxivgate is that it’s going to drive everyone to use AIs to ensure their papers are in compliance because AIs are literally the only entity with the patience to read an entire paper.
14
82
4
1,424
40,892
Out of all GenAI tools Suno remains my favorite. My latest favorite thing is feeding it small compositions, no matter how badly played, and it makes complete songs out of them.
1
44
Näitä asioita voisi myös luotettavasti tutkia. Sitten ei tarvitsisi nojata yhden ihmisen ideologian värittämään kokemusperäiseen anekdotaaliseen todistusaineistoon.
43
Samuli Laato retweeted
the whole "cannot use our models to develop a competing product" is getting really confusing now that model labs are launching regular products like Claude Design
70
53
16
1,919
329,114
Samuli Laato retweeted
Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data. Try it below! with @AlecRad and @status_effects 🧵
195
451
300
3,614
1,469,615
Shoutout to the Ducks organ player for the sick Zelda theme rendition
67
469
33
6,306
352,018