Sabitlenmiş Tweet
I am Gandalf the Green, and I come back to you now, at the turn of the tide
English
Universal Mind
4.3K posts
Universal Mind
@universalmind
Accelerating into the AI future.
Katılım Mayıs 2008
476 Takip Edilen2.5K Takipçiler
@chamath Yeah, UniversalContext was built for exactly this reason.
universalcontext.ai
English
There’s something equally important to attorney-client privilege and that is confidentiality in general.
How many white collar workers are currently feverishly uploading decks, models, PDFs and otherwise confidential work product into public LLMs trying to do their job better or otherwise get an edge right now?
Non-zero.
How will this work, then, when a company claims copyright or confidentiality violations because it turns out an employee broke their employee agreement when using one of these tools.
But, then what? Do the tools purge that information? Can you undo any prompt/response meta data? Agent traces?
Of course not.
When technology innovations meet established regulations, regulations usually win.
Corporations will need to adapt to the ruling below and extend its implications beyond A/C Priv to everyday work.
Moish Peltz@mpeltz
Your AI conversations aren't privileged. Yesterday, Judge Jed Rakoff ruled that 31 documents a defendant generated using an AI tool and later shared with his defense attorneys are not protected by attorney-client privilege or work product doctrine. The logic is simple: an AI tool is not an attorney. It has no law license, owes no duty of loyalty, and its terms of service explicitly disclaim any attorney-client relationship. Sharing case details with an AI platform is legally no different from talking through your legal situation with a friend (which is not privileged). You can't fix it after the fact, either. Sending unprivileged documents to your lawyer doesn't retroactively make them privileged. That's been settled law for years. It just hadn't been tested with AI until now. And here's what really hurt the defendant: the AI provider's privacy policy (Claude), in effect when he used the tool, expressly permits disclosure of user prompts and outputs to governmental authorities. There was no reasonable expectation of confidentiality. The core problem is the gap between how people experience AI and what's actually happening. The conversational interface feels private. It feels like talking to an advisor. But unless you negotiate for an enterprise agreement that says otherwise, you're inputting information into a third-party commercial platform that retains your data and reserves broad rights to disclose it. Judge Rakoff also flagged an interesting wrinkle: the defendant reportedly fed information from his attorneys into the AI tool. If prosecutors try to use these documents at trial, defense counsel could become a fact witness, potentially forcing a mistrial. Winning on privilege doesn't make the evidentiary picture simple. For anyone advising clients or managing legal risk, this is a wake-up call. AI tools are not a safe space for clients to process their counsel's advice and to regurgitate their legal strategy. Every prompt is a potential disclosure. Every output is a potentially discoverable document. So what do we do about it? First, attorneys need to be proactive. Advise clients explicitly that anything they put into an AI tool may be discoverable and is almost certainly not privileged. Put it in your engagement letters. Make it part of onboarding. Don't assume clients understand this, because most don't. Second, if clients want to use AI to help process legal issues (and they clearly will, increasingly), then let's give them a way to do it inside the privilege. Collaborative AI workspaces shared between attorney and client, where the AI interaction happens under counsel's direction and within the attorney-client relationship, can change the analysis entirely. I'm excited to be planning this kind of approach, and I think it's where the industry needs to head. storage.courtlistener.com/recap/gov.usco…
English
@theallinpod When the people from Universal Mind created UniversalContext they didn't mean to wipe away trillions in SaaS value. It's just working out that way.
universalcontext.ai
English
POD UP! 🚨
-- Epstein Files
-- Is SaaS dead? Software stocks crash out
-- Moltbook panic
-- SpaceX-xAI merger
-- Trump's Fed pick
(0:00) Besties intros: Brad Gerstner joins the show
(3:16) Epstein Files
(15:45) SaaS stocks crash out
(35:11) Moltbook panic
(47:37) Trump selects Kevin Warsh as new Fed Chair
(1:00:50) SpaceX and xAI merge
(1:10:45) Brad's major win with Trump Accounts
English
We see product-market fit
x.com/GrecoGum/statu…
Greco Gum@GrecoGum
There are golden ratios everywhere for those with the eyes to see
English
GPT-5 turned out to be surprisingly strong in Cursor. Excited for Codex
Sam Altman@sama
GPT-5-Codex is here: a version of GPT-5 better at agentic coding. It is faster, smarter, and has new capabilities. Let us know what you think! The team has been absolutely cooking, very fun to watch.
English
Different view on how effective AI agent efforts have been. PwC reports 53% of US firms now deploying them in IT & cybersecurity for massive efficiency gains.
A bit contrary to the MIT study.
pwc.com/us/en/tech-eff…
English
@mattturck Yeah.... fully utilizing YouTube training data is a huge advantage.
Will be interesting to see how they attack Anthropic's AI marketshare in the Enterprise.
English
Somehow Google went from being perceived as an AI loser a year or two ago to releasing the most exciting AI products in 2025 - Veo 3, Genie 3, and now 🍌. Remarkable.
English
Nice.
Here is a really good, detailed overview of Software Factory from an early Alpha user.
We’re on the right track.
Chris Stanchak@chrisstanchak
Just tried 8090 Factory Alpha (cc @chamath). It slows you down to speed you up — forcing clarity, specs, and discrete units of work instead of messy one-shot builds. Check out my review...
English
We read the MIT study showing a 95% failure rate for AI projects. This was not a surprise to us. We have seen many huge shifts in technology and this was guaranteed. Here are our thoughts on how to avoid.
universalmind.com/thoughts/95-pe…
English
@sama @yacineMTB Sorry about that whole anti-christ thing... 🤣
English
I AM SO SORRY FOR DOUBTING YOU SAM ALTMAN I AM SO SORRY FOR SAYING THAT YOU WERE THE ANTICHRIST. I DIDNT REALIZE THAT YOUR PLANS WERE MEASURED IN CENTURIES. I HOPE YOU FORGIVE ME FOR EBERYTHING I AM SO GOAD YOU KEPT CONTROL OF OPENAI I AM SO SORRY 😭😭😭😭 PLEASE FORGIVE ME
OpenAI@OpenAI
We released two open-weight reasoning models—gpt-oss-120b and gpt-oss-20b—under an Apache 2.0 license. Developed with open-source community feedback, these models deliver meaningful advancements in both reasoning capabilities & safety. openai.com/index/introduc…
English
Very cool for realtime agents.
Matt Shumer@mattshumer_
It's over. OpenAI just crushed it. We have their o3-level open-source model running on @GroqInc at 500 tokens per second. Watch it build an entire SaaS app in just a few seconds. This is the new standard. Why the hell would you use anything else??
English