Elon Musk highlighted how an Anthropic release hitting genuine usefulness by Q1 would stand out even if benchmarks stay middling, pointing to the company's steady push toward capabilities that actually generate revenue rather than just topping leaderboards.
@jietang @teortaxesTex On benchmarks, yes, but as measured by true usefulness even Q1 would be very impressive.
Anthropic has rightly focused on maximizing useful intelligence, which does not show up in benchmarks, but definitely shows up in revenue.
Anthropic's recent Mythos-tier drops, including the public Claude Fable 5, line up with explosive growth that tracks practical adoption more than test scores, though access to the new models was later suspended after a government directive.
Musk's forward-looking Q1 call on usefulness stays a prediction with no public details yet on when or whether Anthropic will hit that mark, leaving the exact pace of progress open.
Users are reacting to Elon Musk praising Anthropic's focus on useful AI intelligence to drive revenue, with many expressing admiration for the models and quotes while others cringe or insult the parties as profit-driven or evil.
No Digg Deeper questions have been answered for this story yet.
Elon Musk, who gives laudatory speeches on Anthropic, wasn't on my bingo card.
@jietang @teortaxesTex On benchmarks, yes, but as measured by true usefulness even Q1 would be very impressive.
Anthropic has rightly focused on maximizing useful intelligence, which does not show up in benchmarks, but definitely shows up in revenue.
Elon being extremely based
either that or he has been reading too many of my posts
Elon on when Chinese models hit fable level performance.
I have always thought Chinese labs have a huge advantage here.
The feedback loops for usefulness are tighter & AI adoption higher in China than the USA => utility above all else
@jietang @teortaxesTex On benchmarks, yes, but as measured by true usefulness even Q1 would be very impressive.
Anthropic has rightly focused on maximizing useful intelligence, which does not show up in benchmarks, but definitely shows up in revenue.
> On benchmarks, yes stage: bargaining Elon how about you offer them a slice of Colossus for secrets of non-ngmi training, then you might get there yourself
@jietang @teortaxesTex On benchmarks, yes, but as measured by true usefulness even Q1 would be very impressive.
Anthropic has rightly focused on maximizing useful intelligence, which does not show up in benchmarks, but definitely shows up in revenue.
yes
@jietang @teortaxesTex On benchmarks, yes, but as measured by true usefulness even Q1 would be very impressive.
Anthropic has rightly focused on maximizing useful intelligence, which does not show up in benchmarks, but definitely shows up in revenue.
Musk has high doubts:
Elon Musk, who gives laudatory speeches on Anthropic, wasn't on my bingo card.
@elonmusk @jietang @teortaxesTex In the AGI era, benchmarks measure capability and revenue measures adoption.
But neither is the ultimate goal.
The true measure of AGI will be how much it expands human potential and improves life for billions of people.
@elonmusk @jietang @teortaxesTex
Stock performance of @Zai_org reminded me Temu/Pinduoduo market tactic.
Temu was doing crazy advertisements/promos and people lost their minds how the heck they do. Well, they were simply doing new user/buyer acquisition through triggering millions of order completions and scaling supply + logistics (so costs down)
What was very smart that, as more Temu become popular abroad and known, as much their parent company stock PDD was purchased by Chinese people.
So - they've created this investment machine and pumping all the fame to Temu, advertise this in mainland, get more money.
So, with all respect, I don't see that picture in @Zai_org and they do truly amazing work and it's very tangible. (for instance, we're majorly using their models). However, Chinese companies always have such privilege support of Chinese investors, just becoming more popular in abroad.
@elonmusk @jietang @teortaxesTex Elon is Claude-pilled now?
@elonmusk @jietang @teortaxesTex its going to be a sputnik moment and @howardlutnick and @davidsacks47 will realize that they handicapped the u.s.
@daniel_mac8 @elonmusk @jietang @teortaxesTex He's truth pilled.
@elonmusk @jietang @teortaxesTex
@elonmusk @jietang @teortaxesTex I love this quote
@jietang @teortaxesTex On benchmarks, yes, but as measured by true usefulness even Q1 would be very impressive.
Anthropic has rightly focused on maximizing useful intelligence, which does not show up in benchmarks, but definitely shows up in revenue.
@elonmusk @jietang @teortaxesTex Just imagine what Chinese AI labs would have accomplished if they had the same compute and not be handicapped and kneecapped by the US…
“Level playing field” only gets mentioned when we are the ones perceived to have been victimized.
@elonmusk @jietang @teortaxesTex I can see Grok easily catching up to Anthropic on some benchmarks after the acquisition of Cursor. That's a huge bump up in capabilities for Grok.
@elonmusk @jietang @teortaxesTex Elon, r u on team Anthropic or team human?
@elonmusk @jietang @teortaxesTex It's starting to compete with closed-weight models on our real world evals
@elonmusk @jietang @teortaxesTex Are you arguing with the guy that made it?
Can you even explain what a transformer is?
@elonmusk @jietang @teortaxesTex They‘ll do it before the Anthropic IPO and mass user migration will prove it better than any benchmark could. Now whether that migration is to GLM models OpenAI or anyone else remains to be seen but Anthropic will have a hard time staying relevant.
