@kettlecorn

Developer and designer. https://xitter.cf/t.co/BFJAVgEUXo https://xitter.cf/t.co/tP44Plt3Uu https://xitter.cf/t.co/eot0pEir2C

 Philadelphia
ianjk.com
Joined February 2008
  • Tweets5.5K
  • Following2.8K
  • Followers3.8K
538Media
Today I'm open-sourcing Tangle: a new library that automatically 'entangles' and syncs WebAssembly. Play with the demo: tanglesync.com And check out the code: github.com/kettle11/tangle
31
332
36
2,611
302,334
I am making some progress in this space. Presently I have a language that is much slower than Rust, less useful than Go, and a 'type' system that will upset many people. Am I doing this right?
Matt Pocock
@mattpocockuk17h
The next great "language for agents" should be: - As fast as Rust - As little design space as Go - Have the strictest type/verification system ever written Bend seems like an interesting first effort
2
7
796
more replies
I think this is directionally correct. Rather than bolt a language to a type system a type system is treated as one approach to proving things about computation. The prototype compiler is designed to compose its own proofs with proofs from outside sources.
1
56
An aim is to make proofs easily searchable, so if someone asks a 'question' about a snippet of code if the answer has been found it can be quickly reused. I'm not sure when I'll release this prototype, but I sure am spending a huge volume of tokens on it!
50
Programming languages of the future should be designed for greater interoperability regardless of the compiler used. If a better compilation of some code is provably equivalent to the code it should be possible to use it, regardless of source.
Charlie Marsh
@charliermarshOct 24
What happens if these forks become good enough to compile your projects, and are 10x or 100x faster? Do they have staying power?
2
1
6
531
In such a future a language could have a range of compilers for different parts of the program, and for commonly used chunks of code there could even be by-hand targeted compilations allowing an AI or human to create a bespoke compilation.
1
133
The main role of the language and the reference compiler may be to define the language's semantics formally and to create a framework for coordination between various compilers.
105
Another way formal proofs change programming with AI is that they enable confidence in answers from faster / cheaper but less intelligent AIs. We're burning a lot of expensive (and slow) intelligence with the current approaches.
1
3
314
Ian Kettlewell retweeted
Aws Albarghouthi@awsTOOct 23
The trajectory is clear
Anthropic
@AnthropicAIOct 22
New on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here: anthropic.com/research/forma…

LLMs are making formal software verification feasible at a much larger scale

Claude formally proved security properties in the compiled code of 123 open-source libraries. Along the way, it found bugs that had eluded detection for years.

anthropic.com
6
17
1
112
6,413
The other day when I posted about software metamorphosing this is a large part of what I meant. Formal verification will become essential for security and cheap via AI, but it also unlocks fundamentally different capabilities in how software is built & operated at scale.
Anthropic
@AnthropicAIOct 22
New on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here: anthropic.com/research/forma…

LLMs are making formal software verification feasible at a much larger scale

Claude formally proved security properties in the compiled code of 123 open-source libraries. Along the way, it found bugs that had eluded detection for years.

anthropic.com
2
1
3
271
As a 'simple' example if you can formally verify two programs correctly implement the same spec you can freely pick the faster one. Open-source projects could automatically accept faster, smaller, or better code that adheres to a spec.
3
91
The Rust game dev community with ECSs.
Ian JCV@jnvciaOct 21
>open source game engine >ECS come on guys can we try to not follow trends for one minute
1
3
310
Note that the engine linked in the quote-tweet isn't a Rust one. ECSs are very good, but I think they're a tool for a job and not a tool for every job. Some problems map poorly to a general purpose ECS.
97
These forces are coming for all software. If you want people to value your work you must find value beyond the software itself. That can be your unique personality & identity, an enduring social network, brand loyalty, cloud compute, etc.
Andrew Eiche@buddingmonkeyOct 20
This new world changes indie gamedev. If you want to protect your code you can’t use Unity or code that compiles to a VM (IL2CPP won’t be a defense for long) If you want to protect your idea, you need to show off depth of quality content. Simple mechanics will be cloned
1
4
299
On that first point I think for indie devs they can succeed by making things that distinctly reflect the personality & life experiences of themselves, the creators. It's possible to clone code, it's much harder to authentically clone you.
1
62
My mental-model of where software is going is that software is entering its goo phase before transforming into something different, perhaps a butterfly. I think a lot of the industry is building as-if much of the caterpillar is going to persist after the goo phase.
3
1
14
522
Speaking of web standards more people should take a look at custom compression dictionaries are pretty neat: developer.mozilla.org/en-US/…

Compression Dictionary Transport - HTTP | MDN

Compression Dictionary Transport is a way of using a shared compression dictionary to dramatically reduce the transport size of HTTP responses.

developer.mozilla.org
Joost de Valk
@jdevalkOct 16
"we're learning that there's web standards that actually improve how our stuff works” Maybe have a scroll through specification.website

The Website Specification

A platform-agnostic, full specification of the technical features a good website should have. Built in the open under an MIT licence.

specification.website
2
1
9
764
A technique in compression is to identify recurring patterns and assign them a small key. If both sides know the key they can share just the key and expand it to the full value. Brotli does this for common sequences in web files. Custom dictionaries let you define your own.
1
163
As an example if you have a website with many similar web pages you could use this so that mostly only deltas between pages are sent, which is much smaller. Similarly for a game engine with a particular file format you could use this to shave off a lot of transmitted bytes.
1
138
Load more