Ashraf

6.3K posts

Ashraf banner
Ashraf

Ashraf

@ashalhashim

Business Leadership at @anthropicAI . Advisor to a few startups. Happy Baba.

San Francisco Katılım Ocak 2009
630 Takip Edilen1.1K Takipçiler
Ashraf retweetledi
Manouck
Manouck@Manouck44·
On peut critiquer Israël de Paris ou de New York, mais le faire depuis le cœur de Tel-Aviv, en hébreu, donne une tout autre résonance. Shlomo Sand appartient à cette lignée de "Nouveaux Historiens" qui ont décidé de passer les archives nationales au scalpel.
Manouck tweet media
Français
36
744
1.7K
222.9K
Ashraf retweetledi
Amjad Masad
Amjad Masad@amasad·
Right on track. Sonnet 4.5 is a jump in Agentic coding worthy of a major version bump.
Amjad Masad tweet media
Amjad Masad@amasad

In many ways AI progress feels incremental. Remember the GPT-2 to 3 jump or 3 to 4. It was mind-blowing even for pure consumers, who now can’t tell how GPT-5 is an upgrade. BUT there is one capability where progress is clear and measurable: Autonomous agents. In one year Replit Agent, for example, went from being able to do useful work for 2 minutes max to now up to 6 hours! It no longer feels like using software, and more like delegating work to a digital employee. Why is this a big deal? AI will start to transform the economy. Consumer chatbots are limited in terms of economic change—although I don’t think we’ve seen the end of sociological change from it—but the sky is the limit in terms of automating work. I bet by 2027, we’ll see the impact very clearly in macro-economic measures. You wouldn’t have to look for it to see it. However, I also think that the world will not look that much different, and many predictions will look absurd. Reports like “AI 2027” are works of fiction than actual attempt at informing the public. How we work will change, but we can adapt to that very quickly in the same way we adapted to Zoom and then people will look around and ask where is the scifi future we were promised. It’s important for policymakers to get sober and start thinking clearly about the implications for the economy, unemployment, and economic inequality without engaging in fantasy. The change is real, coming, and fairly predictable in terms of what it could look like.

English
29
36
550
57.8K
Ashraf retweetledi
Ashraf retweetledi
Katie Halper
Katie Halper@kthalps·
You may have deleted this tweet but here's a screenshot to document your depravity @DanFriedman81
Katie Halper tweet media
English
750
6.5K
32.7K
801.2K
Ashraf retweetledi
jasper nathaniel
jasper nathaniel@infinite_jaz·
🧵My jaw is still on the fucking floor after reading Jenn Counter’s “Rules of Offensive Influence” — an unapologetic playbook for psychological warfare — that is published on Columbia’s website. I'm going to share some excerpts, then post the whole thing. As you read, keep in mind, Counter serves as both a “Subject Matter Expert” for Safe Reach Solutions — the firm overseeing GHF’s operations in Gaza — and, according to Columbia's website, a “Course Associate” in the Strategic Communications department at the School of Professional Studies.
jasper nathaniel@infinite_jaz

Jenn Counter—a VP at Orbis and “Subject Matter Expert” at Safe Reach Solutions, both central to designing and running the GHF—teaches “Strategic Communications” at @Columbia.

English
108
1.6K
5.3K
541.2K
Ashraf retweetledi
Paul Graham
Paul Graham@paulg·
Jessica: Do think one day people will look back and ask how America could let this this happen? Me: A lot of people are already asking that today.
Paul Graham tweet media
English
354
1.2K
6.5K
396K
Ashraf retweetledi
Paul Graham
Paul Graham@paulg·
Every child in this picture is now dead.
Paul Graham tweet media
English
422
3K
10.9K
1.1M
Ashraf retweetledi
Nassim Nicholas Taleb
Nassim Nicholas Taleb@nntaleb·
53 years ago, Israel assassinated a Palestinian poet, Ghassan Kanafani (& killed a 17 y.o. girl in the process). It was Golda Meir's stated policy of driving Palestinian culture & consciousness to extinction via assassination. "No, it was not a blatant act of terrorism".
English
0
927
3.7K
138.2K
Ashraf retweetledi
Kamell
Kamell@kamellperry_·
Cursor is incentivized to make you spend more, Claude Code is incentivized to make you capable of building more. I don’t want to yap, but when you’re a model provider vs a router, you get to make tools optimized for the model in a way others just can’t. If you’re not using Claude Code you’re already behind.
English
3
2
32
6.3K
Ashraf retweetledi
Tyler
Tyler@TylerDurden·
ZXX
1.4K
6.8K
65.1K
6.2M
Ashraf
Ashraf@ashalhashim·
@amasad Epic, can’t wait to listen
English
0
0
0
9
Ashraf retweetledi
Amjad Masad
Amjad Masad@amasad·
Anthropic deepens their lead on Agentic workloads. Congrats to the team.
Amjad Masad tweet media
English
35
44
968
80.7K
Ashraf retweetledi
Ashraf retweetledi
Amjad Masad
Amjad Masad@amasad·
“Something must change and quickly or the blood of innocents will forever be stapled to the recorded rule of Trump and the MAGA right.”
Amjad Masad tweet media
English
96
143
995
132.6K
Ashraf retweetledi
Alex Albert
Alex Albert@alexalbert__·
We just launched the Max plan for Claude. $200/month, extremely high rate limits (20x pro), priority responses during periods of heavy traffic, and early access to new experimental features. Lots more fun benefits for Max users on the way, stay tuned.
English
199
59
1.4K
224K
Ashraf
Ashraf@ashalhashim·
@skdh Full disclosure I work at Anthropic
English
0
0
0
377
Sabine Hossenfelder
Sabine Hossenfelder@skdh·
I genuinely don't understand why some people are still bullish about LLMs. I use GPT, Grok, Gemini, Mistral etc every day in the hope they'll save me time searching for information and summarizing it. They continue to fabricate links, references, and quotes, like they did from day one. I ask them to give me a source for an alleged quote, I click on the link, it returns a 404 error. I Google for the alleged quote, it doesn't exist. They reference a scientific publication, I look it up, it doesn't exist. Happens all the time. Yes, it has gotten somewhat better in the past 2 years in that with DeepSearch and chains of thought about 50-60% or so of the references exist. By my personal estimate currently GPT 4o DeepResearch is the best one. Grok in particular often doesn't include references even if asked. It can't seem to link even to tweets. It's hugely frustrating. Yes, I have tried Gemini, and actually it was even worse in that it frequently refuses to even search for a source and instead gives me instructions for how to do it myself. Stopped using it for that reason. I also use them for quick estimates for orders of magnitude and they get them wrong all the time. One thing they do save me time with is unit conversion and collecting all kinds of constants. You'd think though that this shouldn't take a 100 million++ LLM to get done. Yesterday I uploaded a paper to GPT to ask it to write a summary and it told me the paper is from 2023, when the header of the PDF clearly says it's from 2025. I don't even know what the heck is going on there, but intelligence ain't it. I sense that a lot of people now think knowledge graphs will fix the LLM-issue, but no, they will not. They cannot. Even in the case that knowledge graphs would prevent logical inconsistency 100%, there are a lot of text-constructions that are perfectly logically consistent but have zero relation to reality. Companies will keep pumping up LLMs until the day a newcomer puts forward a different type of AI model that will swiftly outperform them. On that day, it will become apparent that a lot of companies have been hugely overvalued. It will be a very bad day for the stock market.
English
1.2K
941
6.9K
1.9M
Ashraf
Ashraf@ashalhashim·
Hi @skdh - big fan of yours, you’ve made physics/science so accessible for my children, thank you. Very very curious why you don’t mention Claude, especially Sonnet 3.7 which has a citations feature and is generally regarded as best performing with respect to hallucinations and safety overall. Or why coding isn’t mentioned, as this is the one key area of productivity the market seems to agree today’s LLMs nail.
English
0
0
0
618