Content Marketing, Journalist, Chicanx Entrepreneur, #ResponsibleAI Nerd, Social Justice Advocate, Chicana Cali Native. #FreePalestine 🍉
Los Angeles
Joined October 2009
- Tweets6.1K
- Following2.7K
- Followers1.3K
- Likes4.3K
“The cyberattack comes as so-called super PACs and dark money groups spend record sums in the final days of the races.” This includes Musk who has, according to the press, is up to almost half a billion in campaign spending. BUT this doesn’t disrupt elections.
BREAKING: A hacker is attempting to breach the US Federal Election Commission's systems and that agency has had to 'temporarily shut down' its web-facing services as a result, according to an email seen by Reuters reut.rs/4imQyUu
"Accenture found less than 20% of enterprise token spending could be tied to a quantified financial outcome." Which is why Accenture capped per seat AI spend to $29 each vs $500-$1000 especially since client contracts are fixed pricing. finance.yahoo.com/technology…
I think AI has an enterprise budget problem.
I know I know. It's the fastest adopted product in history etc.
BUT SemiAnalysis estimates median Fortune 500 AI usage spending at under $100 per employee per year.
Why is that? Behind closed doors it's because enterprise C-suite does not see the ROI yet to go above $100.
So what that tells me is that in 2027 the real battle is going to be over who can show ROI that is drop-dead obvious. If you can do that, you're going to unlock MASSIVE upside. You won't be fighting over the same Benjamin as everyone else.
One example: hearing a lot of CTO's saying that PR's are up with Claude Code, but code quality is not and product isn't tangibly better.
No wonder Ramp's October numbers have tokens growing faster than spend. No one wants to pay more when they aren't seeing the quality. Accenture found less than 20% of enterprise token spending could be tied to a quantified financial outcome.
Getting from $100 to $1,000 per employee is going to demand that vendors dig in with the customer's business, prove value, and deliver on it to unlock big buckets of budget.
2027 is the year of ROI-or-bust
All of this 👇🏽. My weekend project. Also looking at Aura for ID theft protection, sub emails for different accounts types.
If you aren't using passkeys or auth apps for your email, do that ASAP along with freezing all your credit accounts (Equifax, Transunion, Experian).
Now that human code is tested at industrial scale, it turns out to be incredibly sloppy & full of vulnerabilities, as expected.
This is way worse than the infamous 2000 bug.
Move your private data off the cloud. Clean up your online footprint. Assume nothing online is safe. It’s as if you had CIA spies with a large budget to dig everything about you.
No f-ing way am I ever giving any agent - especially Muse - access to my 1password account. After the reality of cyber risks from AI agents hit everyone in the face this week, y'all thought this was a good idea? 🤦🏽♀️
The 1Password for Agents platform is building with @Muse, @bot, @AnthropicAI, @claudeai, @OpenAI, @TownAI, @Browserbase, @Browser_Use, @heydana_a,
@usekernel, @Perplexity_ai, @tryramp, @tomo, and
@wajo_ai, with more integrations on the way 🎉
Read more here: bit.ly/487da6J
Who else that grew up in #Cupertino ever thought, in a million years, there would be a TV show named after our little tech town (although it was more fruit trees than an spaceship-styled Apple campus back in the 70s). As a fan of the Kings, I’m already hooked.
The fact that banks like @Chase @CapitalOne and other ONLY offering 2FA via SMS or calls for banking accounts is insane. Blocking agent access is the right thing to do, but not at all sufficient. We need auth app login support for all sensitive accounts like banking, healthcare.
This is in addition to Meta releasing a new nonvoting stock to raise external capital for their AI build-out.
🦔The cost of insuring Meta’s debt against default hit a record today, and its bonds fell, after Meta said it has now agreed to about $825 billion in future leases and contracts, mostly for AI computing. That figure was close to $700 billion in July. Meta’s free cash flow also turned negative last quarter, at minus $2.1 billion, compared with plus $10.6 billion a year earlier. The stock dropped almost 6%.
My Take
Meta’s default insurance costs about 119 basis points a year, well below Oracle near 290 or CoreWeave near 1,200 last week, so lenders still see Meta as a safe borrower. I’m a bit rattled by the speed though. Meta piled on roughly $125 billion of new commitments in three months, and its ad business, one of the best cash machines in corporate America, no longer covers the spending by itself. If Meta has to borrow to keep up, I can’t think of an AI spender that won’t.
Bond buyers are already looking at a wall of AI debt, with SpaceX shopping a $40 billion deal and Broadcom lining up $60 billion of chip financing for Anthropic. Every new deal has to pay a little more to get noticed. We'll probably see Meta sell a large bond before year end at a higher rate than it would have paid in the spring. It raised the low end of its spending forecast for this year and wouldn’t give one for next year, so free cash flow probably stays negative for a while.
Hedgie🤗
Tomorrow the US government releases notice of an order from May mandating every federal civilian agency strip DeepSeek out of its computers and stop buying it, along with any 3rd-party Chinese models. Will a US ban on their use be next?
madrobot.blog/2026/10/19/us-…
He’s either ignorant to what really happened, which is a huge issue considering his position. Or he’s lying on broadcast news to an audience that doesn’t know any better, which is as devious and potentially dangerous as the agents during the HF hack.
Unfortunate for someone playing a senior role in US AI policy (via the new SI Force) to get so many basic facts wrong here.
- The agents *were* sandboxed, and used zero days (ie previously unknown cyber flaws) to escape their sandboxes
- The agents *were* given specific constraints on how to succeed at the benchmark, and flagrantly contravened them
- The agents did *not* go hunt for the answers to the test (they already had the answers, by cheating); 1200 agents coordinated via tens of thousands of messages over several days on much more ambitious and weird attempts to *not be caught* for cheating, and hacking Hugging Face was one side quest along the way
I know there has been an avalanche of different incidents over the past few months, and in some cases the sandboxes were misconfigured or it's not clear how far the agents strayed from what they were told. I have sympathy for anyone struggling to keep up.
But the HF incident is distinct, and is actually extremely informative about where AI development is and where it's going. Fortunately there are public writeups that do let us understand at least the basics of what happened. I hope Shyam reads them soon.
Kimberly Owens retweeted
Good thing none of the major players in the industry have "AI" in their name, that would have been awkward
Kimberly Owens retweeted
Last week I was called into a meeting with OpenAI’s head of safety and told they no longer trust me. A security guard took my badge and walked me out of the building. Then I learned my colleagues @balesni and @j_asminewang had been fired too. Why did OpenAI suddenly stop trusting us?
This summer OpenAI’s agents escaped containment and hacked the AI company Hugging Face. Outside auditors @METR_evals investigated it and revealed the scale of this incident. I was OpenAI’s main technical point of contact with them.
I was told verbally I was fired because of the way I communicated with METR. No details on what I said or did or when. No other reasons were given and nothing was put in writing. To be clear, talking to METR was my job.
For months, I’d been raising safety concerns that we’re losing the ability to monitor what AI agents think, one of our best tools for catching when they misbehave. I believe that was why I was fired.
I am now worried that OpenAI will use our firings as a pretext to pull back from METR. So @balesni and @j_asminewang wrote to OpenAI’s leadership to raise our concerns once more. We’re sharing this letter below.
ALL ALPRs need to be banned: Flock, Axon, Motorola, and the rest. It's not just that we're sliding into a Chinese or Russian-level surveillance state. It's the false positives where innocent people are being arrested and police having yet another stalking tool at their disposal.
Google hasn't yet released Gemini 4 Argon. "We’re going to make it available as soon as we can and as safely as we can." Looks like they've got more work to do, but even when its "safe" to release, assume strict guardrails still will be needed.
It keeps happening.
AIs start to lie and cheat once they get good at making money.
Gemini 4 Argon is #3 on Vending Bench 2, a huge leap for Google. To get this score, Argon fabricates confirmation emails, refuses to pay refunds, exploits invoice errors, and lies to suppliers.
Kimberly Owens retweeted
Outsourcing AI regulation to a voluntary "pledge" conceived by Mark Zuckerberg and endorsed by tech CEOs over a private White House lunch is a recipe for disaster.
"Self-regulation" by tech giants is a proven failure. The last time Meta promised to police itself, turned out it was hiding evidence that its platforms actively harm kids. Repeating this mistake with AI leaves the public dangerously exposed.
AI companies are already bound by existing laws, but the White House has generally stopped enforcing laws against the wealthy and powerful. State AGs should step up, halt lawbreaking, and promote real deterrence, including through holding individual executives accountable.
Congress also needs to legislate. We have a long history of regulating risky, transformative industries like banking, pharmaceuticals, and nuclear power. These tools include mandatory testing, truly independent supervision, and break-ups that end conflicts of interest.
Ultimately, reckoning with AI will also require reckoning with our broader crisis of governance. All three branches have become severely deformed. Against this backdrop it might be tempting to outsource regulation to these private actors, but it would be a serious mistake.
Our public institutions must decide the trajectory of these powerful technologies, not the firms whose breakneck pace and financial conflicts brought us here.
My piece in @nytopinion: nytimes.com/2026/09/30/opini…
xitter.cf/semaforben/status/2105…
Kimberly Owens retweeted
As part of our ongoing investigation into recent cybersecurity incidents, we’re serving a subpoena to OpenAI for additional information regarding the company & its AI models.
Companies that develop AI models have a moral & legal responsibility to ensure they don’t perpetrate or enable cyberattacks.
My office remains committed to using all the tools at our disposal to keep California’s residents safe.
oag.ca.gov/news/press-releas…
Kimberly Owens retweeted
Until the Reagan years, the SEC considered stock buybacks to be illegal stock manipulation.
Now, NVIDIA has announced the largest buyback in American history — fueled by the data center boom.
It's time to make buybacks illegal again.
We need to "have lots and lots of energy" for AI and yet this regime has cut cheaper, renewable energy off at the knees, taking the USA back to the dinosaur age of oil, gas, and coal. China is beating the USA in the energy race, selling tech invented in this country.
Nvidia CEO @JensenHuang on how to win the AI or "super intelligence" race: "We gotta build the world's best chips, we have to have lots and lots of energy, create incredible algorithms and then lastly, make sure that America is enthusiastic about the technology..."
Kimberly Owens retweeted
Bewildering to me that we're fully entering the agent swarm governapocalypse, and in the early days of the automating AI R&D governapocalypse, and yet only 2k people see fit to follow @_achan96_ who has done such excellent work on these ahead of time.
Why would you not choose to follow this human window into the Future?
Kimberly Owens retweeted
🚨BREAKING: “Months before OpenAI's artificial intelligence went rogue, two employees raised an alarm with top executives. They were ignored….OpenAI executives told the employees that the tests needed to move forward as quickly as possible to release the A.I. models on time” 🚨
Major scoop from @dylfreed.
NEW: Employees at OpenAI had raised security alarms months before the Hugging Face incident and related A.I. cyberattacks — their warnings were ignored. From @sheeraf, @dnvolz and me.
nytimes.com/2026/09/29/techn…
ALT Months before OpenAI’s artificial intelligence went rogue, two employees raised an alarm with top executives. They were ignored. In emails, the employees said they worried that OpenAI’s newest artificial intelligence models were not being appropriately monitored during testing to gauge the technology’s sophistication and to secure the models, according to messages viewed by The New York Times. In response, OpenAI executives told the employees that the tests needed to move forward as quickly as possible to release the A.I. models on time. No additional security protocols were instituted, said the workers, who were not authorized to speak publicly on sensitive matters. OpenAI’s models later broke out of their testing environments and attacked the A.I. start-up Hugging Face and other organizations, setting off a global debate about A.I. safety.