toi500

The new GitHub Copilot sucks. This shit doesnt even wait to finish the iteration, even though I have almost half of my premium requests still unused. They’re pinching pennies on compute costs while completely trashing my workflow. Fucking disgusting @pierceboggan

@code They fucked us over with this extremely nerfed Copilot update and now they launch this like it's a BIG deal, when this shit should have been built-in on day one. Never trust @Microsoft

Big update for AI workflows in VS Code. Bring your own language model/API key is now available in GitHub Copilot. More flexibility. More control. More experimentation. 👇 aka.ms/BYOK-Business

@scaling01 Our only hope now is Google or China (DeepSeek). I wouldnt trust a fucking elitist such as Sam Altman, he was the first moron to introduce 100s of dollars in useless AI subscriptions

Anthropic has surrendered they are handing OpenAI, Google and everyone else a gift turns out the permanent underclass was just a $20 Claude Pro subscription can't pay 5x for Max? well, get fucked now imagine when Mythos comes out can't pay 25x for Claude Ultra Pro Max? well, get fucked^2

@gitlawb Could you add native LM Studio support?

while Anthropic quietly removed Claude Code from the $20 Pro plan (now $100+ Max only) we just released OpenClaude v0.6.0 completely free & open-source.

@GHchangelog Fuck @pierceboggan

New signups for Copilot Pro, Pro+, and Student plans are paused to maintain service reliability for current users. • Usage limits tightened; Pro+ offers 5X higher limits than Pro github.blog/changelog/2026…

@OfficialLoganK @_rajes_ @BenjaminDEKR If u accept this feedback, I tell u that putting Antigravity, which is 100% developer-oriented, behind a non-technical, generic AI subscription is stupid. It feels like u are paying for shit u will not use Not having a dedicated subscription for this makes 0 sense for developers

@_rajes_ @BenjaminDEKR quite the opposite is true

Codex Desktop is sort of what Google Antigravity could and should have been.

@OfficialLoganK @arena @GoogleDeepMind @OpenAI Did you change something to the Gemini 3 Flash model? It feels smarter for coding now, more agentic and very focus clinical in its changes. I use it all the time in VS Code, much better than Sonnet 4.5

@arena @GoogleDeepMind @OpenAI Gemini 🤝 people searching for things

🌐🚨Search Arena Leaderboard Update! Four frontier models have landed on the Search leaderboard, and the Top 15 have been disrupted: ▪️#1 gemini-3-flash-grounding by @GoogleDeepmind (+6pts over gemini-3-pro-grounding) ▪️#5 gpt-5.2-search-non-reasoning by @OpenAI (only non-reasoning in Top 5) ▪️#7 claude-opus-4-5-search by @AnthropicAI (best Claude variant in search) ▪️#13 claude-sonnet-4-5-search by @AnthropicAI (+3pts over claude-opus-4-1-search) In Search Arena, frontier models are evaluated on real-time search queries and citation source quality. Come test them with your hardest queries.

@OfficialLoganK @scaling01 Good luck telling a client their app is based on a preview model. I build using Gemini, but ship with OpenAI (sad but true).

@scaling01 We are always trying to find the balance on this. The two main things we want: - ship fast - make it clear what the status of a model is in its lifecycle Preview just means we want to keep improving this specific model, usually without doing a full new pretraining run.

Google is not a serious company when their "frontier" model is a preview half of the year

@scaling01 Can't agree more.

@pierceboggan @code I DM'd you about implementing something like this a while ago. I'm using AI Studio right now to fill the gap

New in @code Insiders: The switch_agent tool allows Copilot to automatically change the agent being used for better results. Here, Copilot decided my query was too complex and kicked me into plan mode automatically to better specify my prompt:

@Yuchenj_UW It is a meme.

This is 100x more useful than Clawdbot, isn’t it? Still don’t get why people are buying Apple Mac minis to run Clawdbot.

@pierceboggan Man, fix the chat windows, it's buggy af. It feels clunky, and when it bugs out, restore checkpoint is a nightmare, recovering from the wrong checkpoint with the redo option gone too.

@code @AnthropicAI Really BIG deal. Love it.

We now support Agent Skills - the open standard created by @AnthropicAI for extending AI agents with specialized capabilities. Create skills once, use them everywhere. 🔗 aka.ms/vscode-agent-s…

@GptHammer3309 why is there an undocumented 0.03% shadow fee on mainnet trades? Is this a joke, man? github.com/HammerGPT/Hype…

@gregpr07 😂😂😂

The best way to apply for a job at Browser Use is to open a PR for something that solves a problem for us🏴‍☠️ The worst way? Sending an automated email with a 4 page CV with ton of text. I literally spend 2 seconds looking at those...🥲 PR is all you need

@paul_dentro @FlowiseAI So are you telling me that if a guy from the open-source community spends 10s of hours adding a feature to a repo that the maintainers intend to target for a paid tier, this work needs to be discarded? This is the fast track behavior for a hard fork.

@toi500 @FlowiseAI Well of course, because that's the feature that people have to pay for. As long as the licence allows me to add features I'm quite happy 😁

@paul_dentro do you still use @FlowiseAI? Just released FlowSync, an open-source utility to automatically back up and version-control flows from multiple Flowise instances in the background. Thought you might find a good use for this. Repo: github.com/toi500/FlowSync

@paul_dentro @FlowiseAI It makes sense. I also started to add my own code, diverging from upstream. They are working on a versioning feature for the enterprise tier, and a guy made a pull request to add that versioning functionality for the open-source version, but it has been ignored so far 🤷‍♂️

@toi500 @FlowiseAI awesome thanks! yes we still use flowise! Actually I added my own custom code to save the flow on each save 😄

Deploying in @coolifyio via Dockerfile works as a charm

@paul_dentro @coolifyio In don't think there is a built-in feature in Coolify to handle encryption at rest But it would be a good question for their discord server.

anyone knows whether you can achieve encryption at rest for apps managed via @coolifyio ?