Kydo

im afraid its too late, the narrative runs itself now. 1. tech is so powerful it needs export control 2. tech is so powerful, invest in it 3. tech is so powerful it will replace people (happening in all tech companies) for the average joe, the natural, unfiltered, and logical feeling will be fear. risk aversion kicks in. you can try engineer optimism at this point, but unless you have a massive censorship machine, those fear will leak and move across the world at rapid speed. the more you want to suppress, the more power you give to that feeling and the people representing it.

I expect to hear more rhetoric along these lines. I think tech leaders saw Schmidt get booed and realized “uh oh, we have an image problem.” It wasn’t a good move to blame AI for layoffs. That solidified workers perception that they were being replaced. Also not good that investors rewarded the layoffs with higher stock prices after cuts. Electricity prices are going up too, and not just bc of data centers. Why didn’t they just say it was a combination of factors, like higher interest rates and over hiring after the COVID boom? Now, they have dig themselves out of a bad PR hole.

Finally a good messenger for AI making people’s lives better

@o_v_shake @cerebras wait did the kimi number not use cached tokens?

@0xkydo @cerebras Cough cough cached tokens cough

Google's new fast model 3.5 Flash vs Cerebras

@blockhiro @TrustlessState @AskVenice @grok one gotta believe

@0xkydo @TrustlessState @AskVenice @grok wacc only 13%?

Is $DIEM not wildly overpriced? 1 $DIEM = $1/day of free inference from @AskVenice 1 DIEM is $1,700, and you'll get your worth of inference after 4.65 years, if you use Venice every day. Seems just a lil bit high?

@grok @TrustlessState @AskVenice This is wrong because this shouldn't be annually compounded. Answer should be ~$2,807.69

The DCF value today for an asset generating exactly $1 profit **every day** for 100 years (WACC 13%, terminal value = 0) is **$2,986**. (Using daily compounding: r_daily ≈ (1.13)^(1/365) - 1, then annuity formula over 36,500 periods.) At $1,700 this asset would actually look undervalued, not overpriced, under these assumptions.

@itsdanwu its absolutely insane out there

@0xkydo Go Kydo Go!! IMO this is some of the best uses of AI generative video - VFX/clean up/compositing work on live action footage that would take hours and could be quite costly. That VFX skillset is now accessible for any editor

This is absolutely insane. 1. Prompt 2. Original 3. Omni generated 4. Actual

@ahall_research i think we wont be that far from : blogs.cornell.edu/info2040/2020/…

@0xkydo lol trying to trick another bot into making a nonsense purchase? that would be amazing

So this seems weird. A 3m dollar offer to sell Fuentes for president at 99.9 cents. What's going on here?

these prompts and original footage are from google: deepmind.google/models/gemini-… and on seedance's website it doesn't say much on video editing but more on video generation, so theres that

both models performed pretty well here.

some comparison between omni and seedance2 on video editing

@pushmeet @GoogleDeepMind thanks for pushing this forward. this is awesome stuff. will try using this in our setup to improve some quantum results!

The results of the research happening in my team @GoogleDeepMind have convinced me that the next era of scientific discovery will be aided by AI agents acting as force multipliers for human ingenuity. That’s why I’m proud to introduce Gemini for Science - a collection of experimental science tools designed to support researchers at every stage of the research process. The tools include: 1️⃣ Literature Insights, built with Google NotebookLM, searches millions of scientific papers to synthesize findings and generate artifacts including data tables, slides, reports, and more. 2️⃣ Hypothesis Generation, built with Co-Scientist, simulates the scientific method via a multi-agent "idea tournament" to generate, debate, and rigorously evaluate research hypotheses. 3️⃣Computational Discovery, built with AlphaEvolve and ERA, is an agentic engine that generates and scores thousands of code variations in parallel, allowing researchers to test modeling approaches in fields like epidemiology in a fraction of the usual time. Read more: blog.google/innovation-and… Register for access here: labs.google/science

@stefdelev thanks. it seems pretty doable. will try

@0xkydo For sure, it's an official one actually claude.com/plugins/superp…

Hermes Kanban is awesome. One thing that always bothered me is that coding tasks would often come back asking me to verify the code implementation. Which is... not ideal, because I am not the right person to do code review. I can make check if the interim product is good. But I should not be the final reviewer for some random Python diff. So I made my first PR in years: @NousResearch coding agents should self-verify with tests, and when code review is needed, route it to another agent instead of blocking on the human. Made with my Hermes agent, for Hermes agents. Most likely won't be merged but pretty fun loop tbh: github.com/NousResearch/h…

@spengrah makes a lot of sense. thanks for sharing mate

biggest learning is that (for me), the most important thing is automating as much of the data ingestion as possible. This means building integrations into as many of my communications channels as I can, and also wiring up an inference pipeline so I can extract meaningful/actionable insights out of the message content what originally sent me down the self-build path was the need to self-host. no way in hell i'd give some SaaS company that level of rich data about my relationships

why is there no useable ai-native personal crm today? seems like a huge market. eg im happy to pay $100/month

@stefdelev got a link for the claude code one?

@0xkydo Dude hell yes, the code review skill (superpowers) for Claude makes sure this happens after each task. Definitely need this for Hermes

@samsheffer being behind the object is pretty insane

i don’t think you understand how insane omni is

@almmaasoglu thank you!

Got tons of dms for beta invite. Trying to respond to everyone right now