This is a good explanation of why the gap between open and closed models is larger than it appears in benchmarks. I would add in that current open models are also more fragile than closed: they handle out-of-distribution problems far less well & have lower emergent capabilities.
@garybernhardt Are you sure you haven’t talked to “a friend at Google” or more recently, “a non-technical PM at a fortune 50” who replaced their $1M Saas with a gas town built replacement? Hilariously, “Software factory” also predates Gas Town as a term for massive government software projects
gpt-5.5 prompt for codex seems to have a duplicated line trying to get it to not talk about creatures?
Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user's query.
[...]
Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user's query
gh link:
https://t.co/1LF8FkRaVf
Here is one for critical missing functionality
https://t.co/N9UPPQa4uE
You can't see what commands the agent is running
https://t.co/6JmSYr2hWN
Agent takes hours to do trivial tasks (I've hit this myself)
If Google AI is amazing, it's not in the realm of software dev
I don't know who to believe on this but if the gemini-cli is any indication of Google's ability to execute on their AI initiatives its an absolute dumpster 🔥. I don't doubt there are smart people working on this but... look at the issue tracker. Basic things do not work
So I leave X for a few months to give my blood pressure a rest, come back, make two posts, both fairly innocuous, and the whole world has decided to hate me now. That's just ducky. Love you guys.
You can choose to believe Google's AI PR team and their core AI researchers, or you can believe your friends who actually work there. I've made my choice.
But I want to clear the air on one thing: I am giving every cent of that crypto money to charity, and the only reason it's not done yet is that Bags is making me conduct two hundred and thirty separate individual transactions on my phone to get the money out. I will have more updates for you when it's done. But for God's sake, don't accuse me of pumping it. I described a phenomenon that was happening to me, in real time, ducked out within six calendar days after I saw what was happening, got death threats, and never spoke of it again. I will, once more, when I do the donation.
"BPC-157 is the biggest scam I've ever seen. It does absolutely nothing. There's no redeemable value to it." - @MartinShkreli
"Why are we going backwards? Why don't we go forwards? What is this urge by the Valley — and I blame the Valley — to go backwards through time and space?"
"This is nonsense. This is not science. Science is controlled experiments that are well-done, very carefully documented, and so forth."
Oh damn, I thought this WAS a joke
... but no, LiteLLM *really* was "Secured by Delve" (the company that rubber stamped all of these audits, and seems to have been on the edge of fraudlent auditing, but useless for sure)
And so unspririsingly LiteLLM was compromised, badly
please shut the fuck up i don't even care about the specific thing you're saying i'm just so tired of hearing predictions one after the other telling me what the future is going to be like just please shut the fuck up
@mitchellh This is so cool. Also surprised how many people have opinions on how you should spend your time. Kudos to you for responding with grace. Hope to see more of this on the app
In Phoenix AZ, the sun rose at 6:46am this morning. It set 6:31pm this evening. They are on Permanent Standard Time. They love it. We should all be like Arizona.