Ariel Jankelowitz

Looking to speed up your annotation? Here's my recipe leveraging zero-shot learning saving time + money w/out sacrificing label quality using: @fastdotai GLIP // @Microsoft @Voxel51 @LabelStudioHQ IceVision // @ai_fast_track @huggingface @zenml_io A short 🧵

@jeremyphoward @RobHackneyEsq @flowersslop that's kind of amazing, given the fascinating history.

@RobHackneyEsq @flowersslop And yet still no mention on Wikipedia or nearly any article or book about the history of LLMs or AI. 🤷

Loving the confidence from @OpenAI Codex. Now let's see what it can do.

Economic paradox. Real estate developers need interesting ground floor tenants to drive sales and rentals of premium units. However, the cost of business is making retail activities unsustainable. Where do we go from here?

@Eric21ND @AmpCode @OpenAI 4 team minimum - also ability to blend prop bets with traditional h2h / roto style

@ajankelo @AmpCode @OpenAI What the limitations on Yahoo? I like their platform for the most part.

Discussing with my son on fantasy baseball. Limitations of each platform. Then I said, "Let's just build our own version with @ampcode or @OpenAI." Now my weekend is set. Wild times!

@jobergum thanks! Have you found the dx experience to be that different?

@ajankelo We migrated to codex and claude code because they told us in the post

Two announcement events recently that I simply don't get. Amp (a coding agent tool) declaring coding agents as dead. Linear (a issue tracking tool) declaring issue tracking as dead.

@jobergum If it is ok to share, what primitives did you wind up migrating to?

In the former case, they announced it without any clear path what does it even mean, so we migrated off Amp to primitives closer to the models (codex and claude code)

@C_Sommerfeldt Also very curious on how @nyctaxi was using slack. What are they going to replace it with?

@C_Sommerfeldt Would be great to see this data in relation to spend and impact on operations / services.

Mayor Mamdani's office says city agencies have submitted proposals for how to shave city spending by $1.7B via savings and efficiencies (his prelim budget already accounted for $1.77B in projected savings, though, so this doesn't do anything to address the underlying budget gap).

@TweetBenMax Very interesting problem. This would be an excellent issue to solve with a systems thinking approach given the time delay aspect and feedback loops. Also, maybe the development should be focused on adaptable skills rather than industry specific skills?

Interesting idea here from Kathy Wylde & Greg Morris for the city to create a "Workforce Development Corporation" & finally 'fix' the city's approach to workforce development programming crainsnewyork.com/op-ed/op-ed-yo…

@charliermarsh Wild, congrats! Love your projects!

We've entered into an agreement to join OpenAI as part of the Codex team. I'm incredibly proud of the work we've done so far, incredibly grateful to everyone that's supported us, and incredibly excited to keep building tools that make programming feel different.

Kudos to @nycfuture for a very interesting talk with @NYCComptroller @MarkLevineNYC on NYC finances. Keeping an eye out on the race to drive more revenue for @nycgov.

@NYCMayor @NYC_SBS Back up.

@NYCMayor @NYC_SBS Website is down for main page and this link. fyi.

Send this to a small business who deserves all the funds. Applications are open for the NYC Future Fund: a loan program with affordable, flexible financing for seasonal and growing small businesses. Apply now at nyc.gov/futurefund

@j_g_allen very true. have been ev since 2019. never looked back.

As I was saying…it’s not that Americans didn’t want EVs, it’s just that us car makers were producing sh*t EVs. Once you have a good EV, you will never want to go back to a gas car. (See next tweet in thread if you haven’t been following…)

@xenovacom @MistralAI Amazing, can’t wait to integrate this.

Introducing Voxtral WebGPU: Real-time speech transcription entirely in your browser. This demo runs Voxtral-Mini-4B, a powerful streaming ASR model from @MistralAI, locally on WebGPU. The model supports 13 languages and is capable of <500 ms latency. Fully private. Zero cost.

@terronk @RoKhanna fair.org/press-release/…

@RoKhanna Pat Buchanan also wrote in his book that America should not have entered WW2 (even to stop the Holocaust). There’s a common thread in Buchanan’s otherwise curious foreign policy. And there’s a common thread in the company you keep from Piker to the Nazi tattoo guy to Buchanan.

You cheerled us into the war in Iraq. You called Pat Buchanan antisemitic for criticizing Israel's bombing of Lebanon in 2006. You are one of the loudest supporters for Netanyahu's war in Gaza. Your neocon ideology has done so much damage to America and people are on to it.

@jackcalifano Not a lawyer but there might also be an issue with New York's ABC Law § 64(7). Which is the 200 / 500 ft rule.

@dex_eve Highly recommend for them to check out Critical thinking for strategic intelligence by Katherine and Randy Pherson.

These are kinda neat tools, but data collection is not intelligence. These people have an "underpants -> ??? -> profit" vision for how the intelligence process works. And without any sort of clear analytical process, you're just producing slop.

One great use case of @AmpCode will be reviving abandoned or orphaned open source projects. Gives you even more confidence with relying OSS.