product @ https://xitter.cf/t.co/gCaNXu1Npc // @datadoghq @optimizely @smartbear @onramp_ai
Boston, Massachusetts
Joined December 2007
- Tweets2.5K
- Following1.5K
- Followers367
- Likes5.8K
opus 5.5 is the first model that feels “there” for more than half of tasks. i still have to steer it 25% of the time, but im imagining in 3-6 months when Sonnet will be equivalent and it is hard to imagine
what's a better setup?
Option A) billions of dollars in compute, storage, and memory sitting on people's desk at home utilized less than 50% of the time
or
Option B) 100% utilization via 10gig symmetrical fiber to every home in america at 1ms latency + 1gig symmetrical satellite with desert datacenters powered by solar
humans are so inefficient
Tyler Mills retweeted
>be diogo almeida
>grow up in the philippines
>win bronze at international math olympiad
>study at rensselaer polytechnic institute
>join google brain
>quit in 2018
>spend 18 months traveling, meditating, playing league of legends
reach top 30 in league
>decide building AI is more fun
>join openai in 2020
>help pioneer RLHF and InstructGPT
>help build chatgpt
>realize LLMs still struggle with reliable decisions
>leave to start your own company (raising 40m seed)
>spend two years building in stealth
launch a new kind of AI model
>3 weeks later raise $870 million series A at a $7.5 billion valuation
not bad for a league of legends player
launched a model
accidentally served trillions of tokens a day
29.4% of the fortune 500 showed up
raised a really big series A from @a16z
it's been 3 weeks
we would like to sleep now
no one wants agents or assistants or ides
they want a solver of problems
they want a connector to people
they want a portal to pleasure
Tyler Mills retweeted
Introducing Eleven v4 and Eleven v4 Turbo, our fastest and most emotive voice models yet.
Ranked #1 by Artificial Analysis.
ux pain in @openwhispr
was using trial, spent 10 minutes dictating something and then it said 'hit limits'
why not fail all future ones rather than lose my work? no warning either
i'm still considering upgrading because i do like it vs @WisprFlow and @WillowVoiceAI but man why are all of their ux so painful?
We brought @typesafeai Jev into Apache DataFusion; unlocking tons of new use cases.
github.com/parable-work/jev-…
hey @instinct @Muse @bot
when you guys figure out how to make credentials not a fucking nightmare let me know
the entire point of agents is pointless if i can’t just set spend controls and tell them to figure out things
constantly being prompted for a 2fa token that times out by the time i get it to them is horrible ux
otherwise im gonna build this myself
Tyler Mills retweeted
I feel like I'm taking crazy pills with how many smart people I read and respect are saying stuff like this.
The idea that the AI labs are suddenly, only now, just in September of 2026, advocating for AI safety, and that they just came around to this position bc of sudden financial precarity, is just completely, utterly, conclusively, 100% wrong.
Here's Dario Amodei telling Ross Douthat in February that he agrees with the case for slowing down; that he's in favor of "collaborating" internationally to organize a slowdown; that "I would be all for" a global slowdown. This was 7 months ago, when Anthropic's annualized recurring revenue was rising faster than any company in modern history. He's been saying stuff like this for years, when Anthropic was worth millions of dollars and when Anthropic was projected to IPO for trillions.
fwiw I think the actual reason leaders of AI frontier labs suddenly seem to agree to pace AI development is
(a) safety measures are currently so crap that they're likely to end up in court if not jail, and they all agree that no one wants that
(b) they see no major new model advancement coming up soon anyway and need an excuse
(c) shift in public opinion
the talk of existential threat is mostly there to keep you distracted and the stock market happy.
Tyler Mills retweeted
I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.
Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Tyler Mills retweeted
from Astra blog - Codex compaction is already great, never think about it, have multi-week threads
turns out avoiding destructive (unrecoverable) summarization actions + letting a very smart model search over previous context…just works
scaling + improving general search works, bitter lesson is inescapable
hot take — the datacenter boom may result in more optimal distribution of compute and investment in internet infrastructure (aka fiber/6g, Wifi 8)
it makes way more sense to have a 10 gig symmetrical connection to a datacenter that can serve ai, gaming, etc than to pay for something that is utilized less than 10% of the time in my home