Harald Klinke<p>As expected: OpenAI just launched GPT-5 for coding and agentic workflows.<br>It outperforms previous models on SWE-bench, Aider polyglot and τ2-bench telecom—handling tool use, bug fixing, and front-end tasks with ease.</p><p>Three sizes (gpt-5, mini, nano), new parameters (verbosity, reasoning_effort=minimal) and support for plaintext-based tool calls via custom grammars.<br>More: <a href="https://platform.openai.com/docs/guides/gpt-5" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">platform.openai.com/docs/guide</span><span class="invisible">s/gpt-5</span></a></p><p><a href="https://det.social/tags/GPT5" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GPT5</span></a> <a href="https://det.social/tags/AIagents" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIagents</span></a> <a href="https://det.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a></p>
eupolicy.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
This Mastodon server is a friendly and respectful discussion space for people working in areas related to EU policy. When you request to create an account, please tell us something about you.
Administered by:
Server stats:
192active users
eupolicy.social: About · Profiles directory · Privacy policy
Mastodon: About · Get the app · Keyboard shortcuts · View source code · v4.4.1
#codegeneration
0 posts · 0 participants · 0 posts today
HGPU group<p>NPUEval: Optimizing NPU Kernels with LLMs and Open Source Compilers</p><p><a href="https://mast.hpc.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://mast.hpc.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://mast.hpc.social/tags/NPU" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NPU</span></a></p><p><a href="https://hgpu.org/?p=30071" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">hgpu.org/?p=30071</span><span class="invisible"></span></a></p>
Harald Klinke<p>Die Grundidee bleibt faszinierend:<br>Eine agentische KI, die ihre eigene Codebasis kennt, strukturiert vorgeht, Rückmeldungen aus der Code-Ausführung einbezieht und dadurch komplexen Code Schritt für Schritt erzeugt.</p><p>Nur leider:<br>Am Ende läuft’s nicht.<br>Und die KI wirkt mit jedem Schritt ein kleines bisschen dümmer. </p><p><a href="https://det.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://det.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://det.social/tags/AIDev" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIDev</span></a> <a href="https://det.social/tags/FailFast" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FailFast</span></a> <a href="https://det.social/tags/PromptJourney" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>PromptJourney</span></a> <a href="https://det.social/tags/DeveloperReality" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>DeveloperReality</span></a></p>
HGPU group<p>ParEval-Repo: A Benchmark Suite for Evaluating LLMs with Repository-level HPC Translation Tasks</p><p><a href="https://mast.hpc.social/tags/CUDA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CUDA</span></a> <a href="https://mast.hpc.social/tags/OpenMP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenMP</span></a> <a href="https://mast.hpc.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://mast.hpc.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://mast.hpc.social/tags/Benchmarking" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Benchmarking</span></a> <a href="https://mast.hpc.social/tags/Package" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Package</span></a></p><p><a href="https://hgpu.org/?p=30005" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">hgpu.org/?p=30005</span><span class="invisible"></span></a></p>
HGPU group<p>P4OMP: Retrieval-Augmented Prompting for OpenMP Parallelism in Serial Code</p><p><a href="https://mast.hpc.social/tags/OpenMP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenMP</span></a> <a href="https://mast.hpc.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://mast.hpc.social/tags/HPC" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>HPC</span></a> <a href="https://mast.hpc.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a></p><p><a href="https://hgpu.org/?p=30004" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">hgpu.org/?p=30004</span><span class="invisible"></span></a></p>
InfoQ<p>Anthropic has upgraded <a href="https://techhub.social/tags/Claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Claude</span></a> – now you can build, host & share AI apps directly from text prompts!</p><p>The new Artifacts feature lets you create functional tools like data analyzers, flashcard generators, or study aids by simply describing the ideas.</p><p>Details here 👉 <a href="https://bit.ly/3ZZDOuT" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">bit.ly/3ZZDOuT</span><span class="invisible"></span></a> </p><p><a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://techhub.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://techhub.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://techhub.social/tags/InfoQ" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>InfoQ</span></a></p>
InfoQ<p><a href="https://techhub.social/tags/CaseStudy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CaseStudy</span></a> - <a href="https://techhub.social/tags/Slack" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Slack</span></a>’s test migration journey: </p><p>20,000 tests, 10 months, one key insight - <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> alone was insufficient.</p><p>In this <a href="https://techhub.social/tags/InfoQ" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>InfoQ</span></a> <a href="https://techhub.social/tags/podcast" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>podcast</span></a>, Sergii Gorbachov, Staff Engineer at Slack, explains why human oversight & conventional tools were essential.</p><p>🎧 Listen now: <a href="https://bit.ly/3FT75Ao" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">bit.ly/3FT75Ao</span><span class="invisible"></span></a> </p><p>📄 <a href="https://techhub.social/tags/transcript" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transcript</span></a> included</p><p><a href="https://techhub.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://techhub.social/tags/Testing" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Testing</span></a> <a href="https://techhub.social/tags/Migration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Migration</span></a></p>
WYM Mod<p>Jonathan Raa reveals that Anthropic, an AI startup backed by Alphabet and Amazon, has reached around $3 billion in annualized revenue, showcasing remarkable growth from $1 billion in December 2024. This rapid expansion signals increasing demand for generative AI, particularly in code generation, positioning Anthropic as a leading SaaS contender against traditional players. Read more about Anthropic's impressive trajectory: <a href="https://www.cnbc.com/2025/05/30/anthropic-hits-3-billion-in-annualized-revenue-on-business-demand-for-ai.html" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">cnbc.com/2025/05/30/anthropic-</span><span class="invisible">hits-3-billion-in-annualized-revenue-on-business-demand-for-ai.html</span></a> <br><a href="https://mastodon.vtip.me/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ArtificialIntelligence</span></a>, <a href="https://mastodon.vtip.me/tags/SaaS" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SaaS</span></a>, <a href="https://mastodon.vtip.me/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a>, <a href="https://mastodon.vtip.me/tags/OpenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenAI</span></a>, <a href="https://mastodon.vtip.me/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a>, <a href="https://mastodon.vtip.me/tags/TechNews" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TechNews</span></a></p>
Christoffer S.<p>For those of you going somewhat in on the using LLM in coding. I'm asking genuinely. I've tried some code generation, but what I struggle with is that code generated may contain methods that dont exists... or language features that have been deprecated etc.</p><p>How do you ensure that generated code is using the latest available code and methods from a library?</p><p>How do you ensure that generated code is NOT in fact using deprecated methods, functions and such.</p><p><a href="https://swecyb.com/tags/Programming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Programming</span></a> <a href="https://swecyb.com/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://swecyb.com/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://swecyb.com/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://swecyb.com/tags/GhostMethods" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GhostMethods</span></a></p>
Karsten Schmidt<p>To put the "large" package size a little more into perspective: I don't know of any other feature-comparable JS vector library which provides all of the following:</p><p>- Generic n-dimensional float, int, uint, boolean vectors<br>- Size optimized versions for 2D/3D/4D (all types)<br>- Multiple-dispatch wrappers (auto-delegating to available optimized versions)<br>- Memory-mapped vectors and optimized versions for various memory layouts (e.g. SOA/AOS)<br>- Optimized versions of many vector-scalar ops<br>- Optimized compound operations (like multiply-add etc.)<br>- Vector randomizations (several approaches)<br>- 99% of GLSL vector operations & conversions<br>- Vector versions of most of JS `Math` ops<br>- Vector interpolations (linear, bilinear, cubic, quadratic...)<br>- 10 different distance functions & metrics<br>- Swizzling & vector coercion/extension<br>- Dozens of additional graphics, statistics & ML-related operations</p><p><a href="https://mastodon.thi.ng/tags/ThingUmbrella" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ThingUmbrella</span></a> <a href="https://mastodon.thi.ng/tags/TypeScript" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TypeScript</span></a> <a href="https://mastodon.thi.ng/tags/JavaScript" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>JavaScript</span></a> <a href="https://mastodon.thi.ng/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://mastodon.thi.ng/tags/Vectors" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Vectors</span></a> <a href="https://mastodon.thi.ng/tags/OpenSource" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenSource</span></a></p>
Karsten Schmidt<p>Just a quick <a href="https://mastodon.thi.ng/tags/ThingUmbrella" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ThingUmbrella</span></a> update to say that I've already replaced the <a href="https://thi.ng/vectors" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">thi.ng/vectors</span><span class="invisible"></span></a> package on the develop branch and after LOTS of deep experimentation have decided NOT to split up the package. There will be a few (minor) breaking changes, mainly because of enforcing more consistent naming and more granularity in some source files (therefore possibly changed imports, though only if you use direct ones for individual functions...). All in all, I've managed to keep the impact on users to a bare minimum (likely unnoticeable for most), even though it's pretty much a complete rewrite of the entire package (with all its ~900 functions)... This package is now almost 10 years old and I'm very happy how this refactor turned out!</p><p>In terms of file size impact: The FULL minified pkg bundle is now 56.4KB vs previously 48.5KB, however the code density has improved and the brotli-compressed pkg size is only 15.1KB (only 1KB larger than before), which I found absolutely incredible! 🎉 I also have to state once more that this package (and most others in <a href="https://mastodon.thi.ng/tags/ThingUmbrella" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ThingUmbrella</span></a>) are _designed for tree shaking_ and bundling. Hardly any project would ever use the full set of functions provided here all at once, most will only use a small/tiny subset...</p><p>Also — more importantly — many of the 185 example projects in the repo are now showing between 2-25% smaller final bundle sizes. Some also have become slightly larger, but so far I found the most by only ~2%...</p><p>Related to this change: I've also updated the <a href="https://thi.ng/color" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">thi.ng/color</span><span class="invisible"></span></a> & <a href="https://thi.ng/matrices" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">thi.ng/matrices</span><span class="invisible"></span></a> packages to be free from dynamic code generation now! The only packages still using `new Function(...)` are the following, but for those it's unavoidable and dynamic code generation is a core feature:</p><p>- <a href="https://thi.ng/pixel" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">thi.ng/pixel</span><span class="invisible"></span></a> (custom pixel format definition/compilation)<br>- <a href="https://thi.ng/pixel-convolve" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">thi.ng/pixel-convolve</span><span class="invisible"></span></a> (custom image convolution kernel compilation)<br>- <a href="https://thi.ng/shader-ast-js" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">thi.ng/shader-ast-js</span><span class="invisible"></span></a> (Shader AST to JavaScript compilation)</p><p>I will do more testing over the coming days, then release new version(s) ASAP...</p><p><a href="https://mastodon.thi.ng/tags/TypeScript" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TypeScript</span></a> <a href="https://mastodon.thi.ng/tags/JavaScript" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>JavaScript</span></a> <a href="https://mastodon.thi.ng/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a> <a href="https://mastodon.thi.ng/tags/Vectors" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Vectors</span></a> <a href="https://mastodon.thi.ng/tags/OpenSource" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenSource</span></a></p>
InfoQ<p>Say hello to Codestral - Mistral AI’s first code-focused <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> model.</p><p>Codestral helps the developers with coding tasks offering efficiency and accuracy in code generation.</p><p>More details on <a href="https://techhub.social/tags/InfoQ" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>InfoQ</span></a> ➡️ <a href="https://bit.ly/4bZD0do" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">bit.ly/4bZD0do</span><span class="invisible"></span></a> </p><p><a href="https://techhub.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a><br><a href="https://techhub.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a></p>
Dirk-Jan Swagerman<p>Mostly, software interfaces are only defined by their signature and without a formal description of the admissible behavior and timing assumptions.</p><p><a href="https://systems.social/tags/ComMA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ComMA</span></a> provides a family of domain-specific languages that integrate existing techniques from formal behavioral and time modeling and is easily extensible.</p><p><a href="https://youtu.be/-bbJTg7pJ-k" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">youtu.be/-bbJTg7pJ-k</span><span class="invisible"></span></a></p><p><a href="https://systems.social/tags/SoftwareEngineering" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SoftwareEngineering</span></a><br><a href="https://systems.social/tags/Interfaces" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Interfaces</span></a><br><a href="https://systems.social/tags/Modelling" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Modelling</span></a><br><a href="https://systems.social/tags/ModelChecking" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ModelChecking</span></a><br><a href="https://systems.social/tags/CodeGeneration" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CodeGeneration</span></a></p>
TrendingLive feeds
Mastodon is the best way to keep up with what's happening.
Follow anyone across the fediverse and see it all in chronological order. No algorithms, ads, or clickbait in sight.
Create accountLoginDrag & drop to upload