Top stories today:
- Trump escalates EU tariff threat after targeting China, Canada, Mexico
- OpenAI unveils Deep Research, o3-powered AI agent
- Sam Altman says OpenAI is on āwrong side of historyā in open-sourcing
- DeepSeekās AI cost $1.6B, far exceeding claimed $6M, research shows
- ChatGPT WAUs, subscribers 3xād to 350M, 15.5M in 2024
0. Data and calendar
All values as of 6 AM ET / 3 AM PT, other than S&P500 and NASDAQ close (4 PM ET / 1 PM PT).
All times are ET.
Listen to our AI-generated podcast summarizing todayās newsletter (beware of hallucinations):
ERRATUM: The headline āU.S. startups raising ā„$1M post-seed -75% to 62 in 2024ā misrepresented the data, as 2023-2024 cohorts are still too early to assess for post-seed trends. h/t to our reader, J.M.
1. Trump escalates EU tariff threat after targeting China, Canada, Mexico
Reiterated a warning to the European Union that tariffs āwill definitely happen.ā
The euro dropped below $1.02 to a 27-month low following Trumpās renewed tariff threats
S&P futures are -1.4%, BTC -6% (since Sat.) to $96K, likely already pricing in high chances of a reversal.
Trump invoked the IEEPA on Sat. to impose a 25% tariff on Canada and Mexico and 10% on China
12:01 AM on Tue.: The measures will come into effect.
Canada announced 25% retaliatory tariffs starting Tuesday, as Mexico vowed to hit back.
China is weighing offering to increase purchases of American goods by $200B over 2 years.
Tariffs lead to inflation, most economists agree.
3. OpenAI unveils Deep Research, o3-powered AI agent for generating comprehensive reports
Accomplishes in-depth research in 5 to 30 minutes by finding, analyzing, and synthesizing āhundreds of online sources.ā
Altman: this can do āa single-digit percentage of all economically valuable tasks in the world.ā
Available for Pro users in 100 queries/ month limit.
Same name as Googleās tool released in Dec.
Deep Research outperforms all models in the Humanityās Last Exam benchmark, slightly ahead of o3ās preview (25.7)
Separately, o3-mini has debuted as OpenAIās new small reasoning model
$4.40/M output tokens, vs. DeepSeekās $2.19/M.
o3-mini medium scored 89 on the Artificial Analysis Quality Index, behind o1
o3-mini lags behind o1 and GPT-4o in agentic tasks and on MLE-Bench, per OpenAI
The model also underperformed on multilingual capabilities
4. Sam Altman says OpenAI is on āwrong side of historyā in open-sourcing tech
