Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
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:

213
active users

eupolicy.social: AboutProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.8

#codeassistant

0 posts0 participants0 posts today
Wildeng<p>Some steps forward in trying to have a personal <a href="https://ruby.social/tags/CodeAssistant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CodeAssistant</span></a> that runs locally, using pre trained specific open source <a href="https://ruby.social/tags/models" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>models</span></a> <br>More info in the readme here <a href="https://codeberg.org/wildeng/ruby_code_agent" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">codeberg.org/wildeng/ruby_code</span><span class="invisible">_agent</span></a></p><p><a href="https://ruby.social/tags/generativeAI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>generativeAI</span></a> <a href="https://ruby.social/tags/Lunarvim" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Lunarvim</span></a> <a href="https://ruby.social/tags/Neovim" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Neovim</span></a></p>
Wildeng<p>Da ingegnere mi sono chiesto il perché ed il percome di molte cose e mi è sempre piaciuto sperimentare. In questo caso mi sono chiesto: è possibile, partendo da modelli open source, crearsi il proprio code assistant personale? Beh non è perfetto ma ci sono vicino:</p><p><a href="https://codeberg.org/wildeng/ruby_code_agent" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">codeberg.org/wildeng/ruby_code</span><span class="invisible">_agent</span></a></p><p><a href="https://livellosegreto.it/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>llm</span></a> <a href="https://livellosegreto.it/tags/aiagent" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>aiagent</span></a> <a href="https://livellosegreto.it/tags/codeassistant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>codeassistant</span></a></p>
Oskar<p>A colleague came across <a href="https://github.com/BerriAI/litellm/blob/main/model_prices_and_context_window.json" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">github.com/BerriAI/litellm/blo</span><span class="invisible">b/main/model_prices_and_context_window.json</span></a> and shared it with me, thought we could maybe convert to a row based format and use in a spreadsheet.</p><p>10 minutes later we instead had this, created in 3 tries by Claude 3.5 Sonnet (new aka 20241022). It was harder deciding how to host it than to "code" it!</p><p><a href="https://austegard.com/llms.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="">austegard.com/llms.html</span><span class="invisible"></span></a></p><p>A searchable, sortable table of 647 LLM models/providers</p><p>How many other such small but previously tedious use-cases are now trivial?</p><p><a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ai</span></a> <a href="https://mastodon.social/tags/CodeAssistant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CodeAssistant</span></a> <a href="https://mastodon.social/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>llm</span></a> <a href="https://mastodon.social/tags/pricing" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>pricing</span></a></p>
Oskar<p>Observation: <br>LLMs like Claude or ChatGPT are *GREAT* for getting you from 0-80%. During this phase you are the product manager and code reviewer, and the LLM is the programmer.</p><p>The last 20% you may want to switch roles: you are the coder and the LLM the reviewer.</p><p>Throughout the process you're best aided by making use of LLM-generated tests, executed by you, in order to keep the two of you from creating regression bugs.</p><p><a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ai</span></a> <a href="https://mastodon.social/tags/claude" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>claude</span></a> <a href="https://mastodon.social/tags/chatgpt" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>chatgpt</span></a> <a href="https://mastodon.social/tags/CodeAssistant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CodeAssistant</span></a></p>
Oskar<p>I would probably not have renamed the function and definitely wouldn't immediately have thought of adding the debounce -- until I inevitably would have ran into an issue later on.</p><p>This going beyond what was asked for is what is delightfully surprising.</p><p>PS! The page in question is here: <a href="https://austegard.com/bookmarklet-installer.html" rel="nofollow noopener noreferrer" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">austegard.com/bookmarklet-inst</span><span class="invisible">aller.html</span></a> </p><p>2/2<br><a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>ai</span></a> <a href="https://mastodon.social/tags/claude" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>claude</span></a> <a href="https://mastodon.social/tags/sonnet" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>sonnet</span></a> <a href="https://mastodon.social/tags/CodeAssistant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CodeAssistant</span></a></p>
Oskar<p>What I'd like to see from any of the big three (four if you consider OpenAI and MS to be separate) here is an interface that manages the state of a work-product _separately_ from the general conversation -- effectively pruning older states of the work product from the conversation history so as to not confuse the model. Though what might be better still would be if the model could manage an initial state plus a change-log (diffs) </p><p><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/CodeAssistant" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CodeAssistant</span></a> <a href="https://mastodon.social/tags/Gemini" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Gemini</span></a></p>
ExploreLive feeds
About