⬢ chaos dancing star @paradigmainc
Joined February 2022
- Tweets 7,903
- Following 388
- Followers 5,291
- Likes 38,978
854 Photos and videos
we're officially live with the open beta of Flywheel.
Flywheel is Paradigma's vision on what autonomous research will run on.
We've built all of this pre-funding, so expect the pace to accelerate significantly.
Super eager to hear your feedback and your ideas.
I'll be live tweeting some of my experiments with Flywheel in a few minutes.
We're just getting started.
11
21
167
25,866
i have a bias here, but the key is that we cannot simply chase and build the next best LLM. we need to build it as a step towards the next big thing, which we can be early on and not inherit cemented practices that the labs, despite their plasticity, inevitably produce. We, as a continent, have INVENTED science, produce the best scientists in the world if normalized by available funding. can anyone guess what the next big thing we’ll use AI for is?
9h
Good points by @pietergaricano against the idea of a European AI lab national champion. If Elon Musk cannot do it, what chance does the French government have? (I am still open-minded about the question, though I have strong priors against.)
open.substack.com/pub/silico…
4
1
20
1,460
14h
Europe is ruled by people who rest on their predecessors laurels thinking that nothing can be done and history is over.
13
448
tensorqt retweeted
The worst thing about "European" tech is that the adjective European is used to excuse subpar performance that cannot compete globally
This is heresy
European should mean the best in the world by far, BTFO'ing all competitors
Don't say "European" if you're not going for best
Jun 13
my take on this is that it’s not even really mistral fault, they actually have lots of clients, and more than a decent amount of gpus and datacenters compared to other open actors
but there is very little (none?) competition in Europe on the « frontier » llm aspect, so they just have to build a « good enough » model, clients choose them because they are European and/or open weight (the fact that enterprises can deploy mistral models on their own infra is a big deal for some companies) so they don’t have to compete on model capabilities with oai and anthropic
12
5
80
6,946
> boomer head of an institution doesn’t believe in agi or something?
FIRE that retard on the spot
> admin has trouble moving things forward and fast?
FIRE them immediately and replace them
we have NO time
6. overrule any decision on nuclear energy and initiate internationally backed reactors paired with datacenters
7. immediately redirect CERN resources, infrastructure and talent towards large scale model training for autonomous scientific discovery
4
30
2,024
incredible
14
828
4. remove anyone older than 45 from decision making positions on frontier problems
5. treat this as a national security threat, override regulation and involve intelligence and military assets as needed
What Europe should do right now:
1. Call all the European researchers working on AI and return them back with same salary (or they can stay but switch career).
2. Fill EU places having GPUs with money, and put those people there.
3. AI partnerships with China India.
6
5
93
7,471
9. launch a single person-led programme fully autonomous and independent from political influence, with large-scale resources and direct accountability on decision-making. The person’s sole focus must be to get us out of this swamp, as soon as possible. The programme must be equivalent to a manhattan project and should be able to quickly and fully access resources and infrastructure as needed
28
514
What Europe should do right now:
1. Call all the European researchers working on AI and return them back with same salary (or they can stay but switch career).
2. Fill EU places having GPUs with money, and put those people there.
3. AI partnerships with China India.
359
361
3,178
199,786
tensorqt retweeted
Jun 13
Its got nothing to do with anthropic and everything to do with the EU being a technologically uninspiring retirement home
Jun 13
Letting EU policy be driven by Anthropic’s press releases and other hype is how one actually loses sovereignty and influence.
1
1
15
512
this is correct. also, this will change at the speed of light. we will pay that and more
Jun 13
it's not lack of compute that's the issze. it's that in Europe, it's unthinkable to pay a guy in his mid 20s $600k salary and give him resources and freedom to train models without having oversight by a committee of gerontocratic professorswho don't keep up with the research
6
2
66
5,830
