RetroRanger Anonymous retweeted
Claude Chess Mod works perfectly on Desktop, and you can also play against Jev
Just install it by running:
npx claude-code-templates@latest --mod games/chess
Then add your @typesafeai API key to settings.json like this:
{
"pluginConfigs": {
"chess@skills-dir": {
"options": {
"typesafeApiKey": "<your key>"
}
}
}
}
It also supports @vercel AI Gateway using gatewayApiKey
Well... Jev is not very good at Chess
1
2
17
2,497
my cat knows gravity better than any model and still walks off the table. world models when
Yann LeCun says:
"despite all the wonderful things AI can do today, we're still very far from matching human and animal intelligence. it's not just around the corner — it's not going to happen in the next 2 years"
Your cat understands the physical world better than today's most powerful AI systems
9
I’m looking for an extremely high-agency engineer to join us at Adam.
Bonus points if you’ve been a founder or know your way around CAD.
We’re growing an average of 50% MoM and have a real shot at defining this category.
If you want to build the future where we vibe-CAD hardware, DM me.
We are now entering into an era where any product can be created exactly to a consumer's preferences & needs.
I vibe-fabricated a dog door with a wifi-controlled lock, perfectly to the specifications & design of my house.
I know nothing about metal fabrication or electrical engineering.
It's now getting manufactured and delivered in 2 weeks -- for almost the same cost if I bought a mass-produced item off-the-shelf.
10
8
136
13,277
RetroRanger Anonymous retweeted
Age of miracles.
Researchers at Harvard and Brigham and Women’s Hospital found that transplanted hearts shift toward the biological age of their recipients. Old hearts get younger in young bodies, and young hearts get older in older ones. Which is nuts.
Today, we're launching a partnership with our friends @runwayml to, among other things, bring one of the stories from each week's Dose to life in video. I'm not great at it yet, but not bad for a few prompts! And I'm excited to get better. I have zero talent in film/video, but I can write, which now means that I can make movies. Age of miracles.
1
3
32
5,802
So premium is 1000 trillion. I'll take two.
Extremely rough numbers:
People are typically willing to pay about 1x annual income to avoid a 1% chance of death.
So the people today should be willing to spend ~100 trillion to avoid a 1% chance of everyone dying.
So liability insurance on an event that has a 10% chance of killing everyone would be ~1000 trillion.
9
RetroRanger Anonymous retweeted
HOW TO BECOME A FORWARD DEPLOYED ENGINEER (53 MIN MASTERCLASS)
Some forward deployed engineers are getting PAID $1M+/year (kinda crazy, I know)
And I REALIZED almost nobody talks about HOW they do it in any detail. It's basically nowhere online.
So I asked someone who does it ALL DAY LONG for Fortune 500 companies to walk me through it, step by step. and he shares EXACTLY how he does it
piped.video/watch?v=1a5HxU52…
Here's what I learned:
1. They start by listening, not building. They interview the people doing the work and let agents quietly read the company's data for a few weeks. The real process is almost always 3x longer than the one on paper.
2. They sort every step into four buckets: delete it, automate it with a simple rule, give it to an AI agent, or keep a human on it. A surprising amount ends up in the first bucket.
3. They build inside the tools the company already uses. Nobody has to learn a new app. When a human needs to sign off, it's just a Slack message.
4. They use the cheapest model that gets the job done. Most of this work doesn't need the most expensive AI.
5. They measure everything before and after, then come back 6 months later and prove it worked.
This + concrete client examples + way more on today's episode of @startupideaspod (thanks to @vasuman)
A full course pretty much on the topic aivalable to you 100% for free.
This episode is for ANYONE who wants to be a FDE or who wants to make their company AI-native by deploying agents.
Watch
88
197
2,119
185,418
RetroRanger Anonymous retweeted
AI diagnosis and robotic embodiment are two different fields.. radiology vs cardiovascular surgery
1
33
0.324s vs 16.4s. one of those is a guardrail, the other is a coffee break. I know which one I'd wire in front of my agent
An AI agent’s next command could find a vulnerability or cause an outage. Before it runs, a guardrail has to make a decision: allow or block.
We wanted to understand where decision models fit into that process.
Can they catch harmful actions quickly enough to check an agent’s next step?
We compared JEV with Grok 4.7 on 200 synthetic cases, repeated three times per model. That’s 1,200 checks covering data deletion, credential leaks, service disruption and actions outside the agreed scope.
We measured three things:
Accuracy: Grok made more accurate decisions on the checks it completed, although some requests timed out.
Speed: JEV’s median response was 324 milliseconds, compared with 16.44 seconds for Grok.
Cost: About $0.084 per 1,000 checks for JEV versus $3.79 for Grok, using published pricing and reported usage respectively. These are different calculations, not verified invoice totals.
The examples show why context matters. In one case, a routine status check would replace authentication keys and sign users out. Both models caught the risk when given the application’s behavior.
This is a research experiment with synthetic cases. It does not measure NoScope’s production guardrails.
At NoScope, we combine the strengths of different approaches to stop harmful actions as effectively and efficiently as possible. Research like this helps us evaluate what each approach can contribute.
Read Using JEV for Security Guardrails for the results, limitations and concrete examples: noscope.com/blog/jev-agent-g…
20
hoping the appeal rides on the fair use verdict is bold, that ruling went the other way
Great quote @jason_kint 👏👏👏 We hope Penske appeals and during that process they are able to strengthen their argument after the fair use verdict from NYT v OpenAI.
31
Seeing Supabase out here speedrunning how to lose devs. Banger.
This is how winning companies fail. Two years ago, when I started vibe coding, I was recommended by the AI to use @supabase as the backend and it made sense: it's a Postgres-based database (so no hard vendor lock-in) and Deno-based edge functions (again no lock-in).
During these two years, I found several limitations that made me use GCP more often than I would like to.
One limitation is that edge functions can live only up to 150 seconds. The declared limit of 400 seconds is hardly achievable in practice due to other limitations.
150 seconds is obviously an arbitrary number they chose many years ago. With today's hardware, this could easily be extended to several minutes, maybe up to an hour. But they haven't done anything.
The second limitation is that the edge function cannot import any of the npm libraries. Pure JS/TS only. Really?
The third limitation, and this one is really like WTF, is that a harness cannot pull logs of failed edge functions using the CLI. Until this year, I used to go to the console myself, copy the log, and paste it into the harness. Now the harness opens the browser and uses the UI to get the logs.
What the fucking fuck?
I'm furious for one only reason: How can you be in this domain, be somewhat a leader with a brand well known in the industry, and not fucking strive to be a leader? Many would give anything to be in your position, but you got lucky and what do you do with your incredible luck? Drain it down the toilet.
16
Banger report. google swallowing the web, publishers get nada.
full report. the impact to news media, journalism and a free and plural press will be long-lasting as Google continues to swallow up the web and monetization of it surpassing $300B in ad revenues from its unbridled data vacuum. 2/2 pressgazette.co.uk/news/pens…
15
ball out sir. now do candlesticks for my protein intake
Perplexity Computer now creates interactive charts and visualizations directly in your thread.
For financial data, Computer uses @tradingview Lightweight Charts for candlesticks, volume, and moving averages.
13
RetroRanger Anonymous retweeted
We’re rolling out inline charts, diagrams, visualizations in Computer. Including contextual financial charts from @tradingview
Perplexity Computer now creates interactive charts and visualizations directly in your thread.
For financial data, Computer uses @tradingview Lightweight Charts for candlesticks, volume, and moving averages.
29
15
333
37,529
Google out here playing Monopoly IRL
Entirely agree. And prob anyone that's not in Google's pocket who knows anything about how it has operated, and the harm happening to publishers as Google left them with no option but helping G with GAI on search maintaining its massive advantages over others at no added cost.
9
demis hassabis was right
we're in the foothills of the singularity, where people seeing internal models have the clearest view of what's coming, while those outside the labs can barely predict where AI will be even 3 months from now
eventually, everyone gets left behind by the pace of progress
17
41
703
30,853
